Indoor path dynamic planning method and system

Abstract

The invention provides an indoor path dynamic planning method and system, and relates to the technical field of path planning, and the method comprises the steps: obtaining a real-time position and a target position of an indoor environment user, and an accumulated path of the user, and extracting accumulated indoor elements in the accumulated path; according to the real-time position and the target position, planning to obtain a plurality of indoor paths, and extracting indoor elements in the plurality of indoor paths to obtain a plurality of indoor element sets; the association degree of the multiple indoor element sets and the accumulated indoor elements is analyzed, and multiple element association degrees are obtained; and acquiring a plurality of congestion parameters of the plurality of indoor paths, configuring congestion weights and element weights according to historical moving paths of the user, calculating to obtain a plurality of path scores by combining a plurality of element association degrees, and screening to obtain an optimized indoor path as a path planning result. The technical problem that in the prior art, a static path planning method cannot provide personalized path planning for a user is solved.

Description

Technical Field

[0001] The present invention relates to the field of path planning, and in particular to a method and system for dynamic indoor path planning. Background Art

[0002] With the popularity of complex indoor buildings such as large commercial complexes and airports, traditional static path planning methods only perform path planning based on a fixed road network topology. This cannot meet the personalized needs of users (such as users preferring to pass through specific shops or avoid stairs), and thus cannot provide personalized path planning for users. Summary of the Invention

[0003] The present invention addresses the technical problem that static path planning methods in the prior art cannot provide users with personalized path planning, and provides an indoor path dynamic planning method and system.

[0004] The technical solution of the present invention to solve the above technical problems is as follows:

[0005] In a first aspect, the present invention provides a method for dynamic indoor path planning, comprising:

[0006] Obtaining the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extracting the cumulative indoor elements within the cumulative path;

[0007] Planning and obtaining a plurality of indoor paths according to the real-time position and the target position, and extracting indoor elements within the plurality of indoor paths to obtain a plurality of indoor element sets;

[0008] Analyzing the correlation between the plurality of indoor element sets and the accumulated indoor element to obtain a plurality of element correlations;

[0009] Multiple congestion parameters of multiple indoor paths are obtained. Based on the user's historical movement path, congestion weights and factor weights are configured. Combined with the correlation between multiple factors, multiple path scores are calculated and selected to obtain the optimized indoor path as the path planning result.

[0010] In a second aspect, the present invention provides an indoor path dynamic planning system, comprising:

[0011] A data acquisition module is used to obtain the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extract the cumulative indoor elements within the cumulative path;

[0012] A path generation module is used to plan and obtain multiple indoor paths based on the real-time position and the target position, and extract indoor elements in the multiple indoor paths to obtain multiple indoor element sets;

[0013] a correlation analysis module, configured to analyze the correlations between the plurality of indoor element sets and the accumulated indoor elements to obtain a plurality of element correlations;

[0014] The optimization output module is used to obtain multiple congestion parameters of multiple indoor paths, configure congestion weights and factor weights based on the user's historical movement path, combine the correlation of multiple factors, calculate multiple path scores, and screen to obtain the optimized indoor path as the path planning result.

[0015] The beneficial effects of the present invention are:

[0016] Compared with the existing technology, this application first obtains the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extracts the cumulative indoor elements in the cumulative path. By obtaining the user's real-time location, target location and historical path information, it can accurately capture the user's recent preference for indoor elements (such as elevators, shops, toilets, etc.) in path selection, upgrading path planning from simple spatial distance planning to intelligent matching based on user personalized needs, and providing key data support for subsequent path optimization. Secondly, based on the real-time location and target location, multiple indoor paths are planned and obtained, and indoor elements in multiple indoor paths are extracted to obtain multiple indoor element sets, which provide reliable data support for subsequent element correlation analysis and path dynamic planning. Thirdly, the correlation between multiple indoor element sets and cumulative indoor elements is analyzed to obtain multiple element correlations. By calculating the correlation between the indoor elements of the candidate path and the user's cumulative indoor elements, the probability that different planned paths meet the user's current path selection preference is quantified, and the user's historical behavior preference is converted into a quantifiable path matching index, making path planning more in line with the user's personal choice preference and providing the necessary data foundation for the final path planning. Finally, multiple congestion parameters of multiple indoor paths are obtained. According to the user's historical movement path, congestion weights and factor weights are configured. Combined with the correlation of multiple factors, multiple path scores are calculated and screened to obtain the optimized indoor path as the path planning result. By analyzing the user's historical movement data, the weight ratio of congestion parameters and factor correlation is dynamically adjusted. Through the weighted scoring mechanism, different paths are comprehensively scored based on congestion parameters and factor correlation. Finally, the optimal path planning result is screened and recommended to users.

[0017] Through the above technical solution, this application fully considers the differences in users' preferences for indoor elements and congestion levels within the path when choosing a path indoors, mines personalized preference characteristics from users' historical data, dynamically adjusts the path planning strategy based on real-time environmental data, and screens and outputs the optimal path planning results to provide users with accurate personalized path planning services. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1A schematic diagram of the process flow of the indoor path dynamic planning method provided by the present invention;

[0019] Figure 2 This is a structural diagram of the indoor path dynamic planning system provided by the present invention.

[0020] In the accompanying drawings, the components represented by the reference numerals are as follows:

[0021] Data collection module 11, path generation module 12, correlation analysis module 13, optimization output module 14. DETAILED DESCRIPTION

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0023] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0024] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0025] Example 1, as Figure 1 As shown, an embodiment of the present invention provides a method for dynamic indoor path planning, including:

[0026] S10: Acquire the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extract the cumulative indoor elements in the cumulative path;

[0027] In indoor environments, users can plan multiple routes from their real-time location to their target location. However, there are significant individual differences in their route selection, particularly in their preferences for specific indoor elements. For example, some users prefer to take elevators rather than escalators or prefer to pass by specific stores. Furthermore, user preferences can fluctuate over short periods of time. For example, shopping routes can change rapidly depending on current needs (such as searching for a specific product). Therefore, dynamic route planning can be performed based on the user's real-time location and target location in the indoor environment, as well as the accumulated indoor elements within the user's cumulative path.

[0028] To address the above issues, this application obtains the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extracts the cumulative indoor elements within the cumulative path.

[0029] Specifically, step S10 in the method includes:

[0030] Obtain the real-time location and target location of indoor users;

[0031] Get the user's path within the preset time range in the past and obtain the cumulative path;

[0032] A preset number of indoor elements closest to the real-time position in the cumulative path are extracted to obtain cumulative indoor elements.

[0033] In the embodiment of the present application, the real-time position and target position of the user in the indoor environment are first obtained. For example, the user's position information in the indoor environment (such as three-dimensional coordinates (X, Y, Z)) can be obtained in real time through indoor positioning technologies such as Bluetooth positioning, GPS positioning, and UWB (ultra-wideband) positioning, and the target position is determined in combination with user input or historical behavior (such as common destinations), and the target position is converted into specific coordinate information, wherein the positioning accuracy needs to reach the meter level (such as 1-3 meters) to ensure the accuracy of subsequent path planning. For example, the real-time position information of the user in the indoor environment (such as (500, 2100, 15.5) and the target position input by the user (such as Shop 1 on the 2nd floor) are obtained through GPS positioning, and the target position (such as Shop 1 on the 2nd floor) is converted into specific coordinate information (such as (300, 1000, 5.5)).

[0034] Secondly, the user's path within the preset time range in the past is obtained to obtain the cumulative path. Specifically, the preset time is set because the user's behavioral preferences in indoor environments have short-term fluctuations. For example, the route taken when shopping may change rapidly with current needs (such as looking for a certain product). The preset time range can ensure the timeliness of the collected data. Setting a shorter preset time (such as 30 minutes) can more accurately reflect the user's current intention, and the user's current preference analysis and behavior prediction can be performed based on this. However, in larger indoor environments (such as airports), since indoor elements such as shops and toilets are far apart, it is not advisable to set the preset time range too short. Therefore, those skilled in the art can flexibly adjust the preset time according to the specific application scenario. For example, in a shopping mall scenario, it can be set to 30 minutes to 2 hours (covering the historical path of the user's current shopping trip), and in an airport scenario, it can be set to 1 hour to 4 hours (covering the user's main activity period from security check to boarding). Furthermore, positioning technology (such as GPS positioning) can be used to continuously record the user's path within a preset time range and arrange it according to the positioning time to form a path sequence consisting of a series of coordinate points, for example, (X1, Y1, Z1) → (X2, Y2, Z2) → ... → (X n , Y n , Z n ), the path sequence is the cumulative path of the user within the preset time range.

[0035] Finally, extract a preset number of indoor elements in the cumulative path that are closest to the real-time position to obtain the cumulative indoor elements, wherein the preset number is to balance the representativeness of the extracted data and data redundancy. If the preset number is too small (such as 2), it may not be able to fully capture the user's current demand preferences. If the number is too large (such as 20), it may include elements that are irrelevant to the current demand (such as a remote corner that the user passed by half an hour ago), resulting in unnecessary waste of resources. This application is based on a large number of experiments and recommends a preset number of 3-5. Those skilled in the art can adjust it according to actual conditions (such as specific application scenarios, flow of people in application scenarios, etc.). Among them, indoor elements refer to the types of environmental entities contained in the cumulative path, such as elevators, stairs, women's clothing stores, service desks, toilets, etc. Each indoor element can be identified by a unique type. Furthermore, extracting the indoor elements closest to the real-time position is to ensure that the user's current selection preferences can be obtained. For example, the preset number is 3, and the three indoor elements closest to the real-time location in the user's cumulative path, such as an elevator, a women's clothing store, and a bathroom, are extracted as the user's cumulative indoor elements. The cumulative indoor elements reflect the user's short-term fluctuating preference for specific indoor elements in the historical movement trajectory. For example, if "women's clothing store" appears in the user's cumulative indoor elements, the user will most likely choose a path with a "women's clothing store" in future path selection.

[0036] In summary, compared to existing technologies, this application obtains the user's real-time location and target location in an indoor environment, as well as the user's cumulative path, and extracts the cumulative indoor elements within the cumulative path. In this way, by obtaining the user's real-time location, target location, and historical path information, it is possible to accurately capture the user's recent preferences for indoor elements (such as elevators, shops, and bathrooms) in path selection, upgrading path planning from simple spatial distance planning to intelligent matching based on the user's personalized needs, and providing key data support for subsequent path optimization.

[0037] S20: planning and obtaining multiple indoor paths according to the real-time position and the target position, and extracting indoor elements in the multiple indoor paths to obtain multiple indoor element sets;

[0038] In indoor environments, users exhibit significant individual differences in their route selection behavior, particularly in their preferences for specific indoor features. For example, some users prefer routes that use elevators over stairs, or routes that pass through specific functional areas like coffee shops and convenience stores. These differences in preferences for indoor features directly influence users' route selection decisions.

[0039] To address the above issues, this application plans and obtains multiple indoor paths based on the real-time location and target location, and extracts indoor elements within the multiple indoor paths to obtain multiple indoor element sets.

[0040] Specifically, step S20 in the method includes:

[0041] Obtain the road network of the indoor environment;

[0042] According to the road network, based on the real-time location and the target location, planning and obtaining multiple indoor paths with the shortest path lengths;

[0043] Indoor elements in the multiple indoor paths are extracted to obtain multiple indoor element sets of the multiple indoor paths.

[0044] In the embodiment of the present application, the road network of the indoor environment is first obtained. For example, the road network can be constructed in the form of nodes and edges, wherein the nodes include large and small nodes, large nodes represent key indoor elements, such as corridor intersections, elevator entrances, stairwells, toilets, etc., and small nodes represent ordinary indoor elements, such as general shops, etc. Each node has precise three-dimensional coordinates and type labels (such as elevators, clothing stores, cake shops, etc.), and edges represent the passable paths between nodes. Weights can be assigned to edges based on physical distance, real-time pedestrian flow, difficulty of passage, etc. For example, in a shopping mall scenario, the nodes of the second-floor road network include escalator entrances, store entrances, toilets, etc., and the edges are corridor paths connecting these nodes, and their weights will be adjusted in real time with the density of pedestrian flow. Furthermore, the road network data can be obtained by digital processing of architectural design drawings, and supplemented with details by manual annotation. At the same time, the road network weights are dynamically updated through environmental data collected in real time by the sensor network (such as real-time pedestrian flow monitored by cameras) to ensure that the road network is consistent with the actual environment.

[0045] Next, based on the road network, multiple indoor paths with the shortest path lengths are planned based on the real-time location and the target location. Specifically, based on the constructed indoor environment road network, multiple shortest paths are planned from the user's real-time location to the target location. The path planning process can be performed using an algorithm, such as the K-short path algorithm, which directly calculates the top K shortest paths from the starting point to the end point. To ensure path diversity, the overlap between multiple candidate paths can be controlled. For example, it is required that at least 30% of the sections of any two paths are non-overlapping to avoid generating overly similar solutions. For example, from the real-time location (e.g., the lobby on the first floor) to the target location (e.g., the cinema on the third floor), the K-short path algorithm generates three paths: Path 1 (a direct elevator path), Path 2 (a two-story escalator path), and Path 3 (a staircase path to the third floor and then to the elevator lobby). These three paths are of similar length but pass through different indoor elements, providing a foundation for subsequent personalized path planning. Furthermore, by pre-calculating candidate paths for high-frequency target locations (such as bathrooms) and caching them, the response time of path generation can be shortened from hundreds of milliseconds to tens of milliseconds, significantly improving the efficiency of path generation.

[0046] Finally, indoor elements within the multiple indoor paths are extracted to obtain multiple indoor element sets for the multiple indoor paths. Specifically, indoor elements are extracted from the generated multiple indoor paths and a set is constructed. During the extraction process, semantic parsing is performed on each indoor element passed by the path, and it is converted into a standardized element type to form an element set. For example, the three indoor paths are: Path 1 (a path directly accessible via an elevator), Path 2 (a path ascending two floors via an escalator), and Path 3 (a path via stairs to the third floor and then to the elevator hall). Indoor elements within each indoor path are extracted separately. For example, Path 1 passes through an elevator, a corridor on the third floor, and a cinema, and its corresponding indoor element set is {elevator, corridor, cinema}; Path 2 passes through an escalator and a cinema, and its corresponding indoor element set is {escalator, cinema}; Path 3 passes through stairs, an elevator hall, and a cinema, and its corresponding indoor element set is {stairs, elevator hall, cinema}. These element sets contain static element information within multiple planned paths, providing multi-dimensional data support for subsequent analysis of user preferences and path element correlation analysis.

[0047] In summary, compared to existing technologies, this application plans multiple indoor paths based on real-time location and target location, extracts indoor elements within these paths, and obtains multiple indoor element sets. This provides reliable data support for subsequent element correlation analysis and dynamic path planning.

[0048] S30: analyzing the correlation between the plurality of indoor element sets and the accumulated indoor element to obtain a plurality of element correlations;

[0049] In indoor environments, when users choose multiple paths from their real-time location to their target location, they are likely to make decisions based on their historical behavioral preferences. For example, within a certain timeframe (e.g., the past hour), the indoor elements they have passed by will influence their current path choice. For example, if a user has passed a "women's clothing store" multiple times in the past hour, they are likely to choose a path containing it in future path choices. Furthermore, the degree of association between different indoor elements varies significantly, as evidenced by their co-occurrence frequency within different paths. For example, because most users prefer to take elevators rather than stairs, the co-occurrence frequency between elevators and stores within a path is significantly higher than the co-occurrence frequency between stairs and clothing stores, reflecting a strong correlation between elevators and stores.

[0050] To address the above issues, the present application analyzes multiple indoor element sets and accumulates the correlations of the indoor elements to obtain the correlations of the multiple elements.

[0051] Specifically, step S30 in the method includes:

[0052] Traversing and selecting indoor elements from multiple indoor element sets, combining the indoor elements with the accumulated indoor elements and inputting them into an element association analyzer, and outputting multiple indoor element association degree sets;

[0053] The mean of the correlation degree sets of multiple indoor elements is calculated to obtain the correlation degrees of multiple elements.

[0054] In an embodiment of the present application, indoor elements are first traversed and selected from multiple indoor element sets, and the indoor elements are combined with the accumulated indoor elements and input into an element association analyzer, and multiple indoor element correlation degree sets are obtained as output, wherein the element association analyzer is constructed based on machine learning. For example, the indoor element set of the second path in the multiple indoor element sets is {escalator, cinema}, and the accumulated indoor elements are {elevator, women's clothing store, bathroom}. Each indoor element in the indoor element set is combined with the accumulated indoor elements in turn, for example, {escalator, elevator}, {escalator, women's clothing store}, {escalator, bathroom}, {cinema, elevator}, {cinema, women's clothing store}, {cinema, bathroom}, and then input into the pre-trained element association analyzer in turn, and multiple indoor element correlation degrees are obtained as output, for example, 0.2, 0.3, 0.6, 0.4, 0.3, 0.6.

[0055] Next, the mean of the multiple indoor element correlation sets is calculated to obtain multiple element correlations. For example, the obtained multiple indoor element correlations (e.g., 0.2, 0.3, 0.6, 0.4, 0.3, 0.6) are output, and the mean is calculated to obtain element correlation = (0.2 + 0.3 + 0.6 + 0.4 + 0.3 + 0.6) / 4 = 0.4, indicating that the current second planned path has a moderate correlation with the user's selection preference. The element correlation quantifies the strength of the correlation between different planned paths and the user's selection preference. A larger element correlation value indicates a higher probability that the planned path meets the user's current path selection preference.

[0056] Furthermore, the training steps of the “element association analyzer” include:

[0057] Based on the indoor path record data in the historical time, the sample cumulative indoor element set and the sample indoor element set are collected, and the indoor element correlation degree of each sample cumulative indoor element and the sample indoor element is marked to obtain the sample indoor element correlation degree set, where the sample indoor element correlation degree includes the probability that the sample cumulative indoor element and the sample indoor element appear in the same indoor path at the same time;

[0058] Build a feature association analyzer based on machine learning;

[0059] The sample accumulated indoor element set, the sample indoor element set and the sample indoor element association degree set are used to perform supervised training on the element association analyzer until convergence.

[0060] In the embodiment of the present application, the training process of the factor association analyzer mainly includes three steps: data collection, model construction, and model training. Specifically:

[0061] First, data collection is performed. Specifically, based on the indoor path record data over a historical period, a sample cumulative indoor element set and a sample indoor element set are collected. The indoor element correlation between each sample cumulative indoor element and the sample indoor element is annotated to obtain a sample indoor element correlation set. The sample indoor element correlation includes the probability that the sample cumulative indoor element and the sample indoor element appear simultaneously on an indoor path, that is, the probability that two indoor elements appear simultaneously on an indoor path. For example, if "elevator" and "coffee shop" appear together 30 times in 100 historical paths, the correlation between "elevator" and "coffee shop" is 30 / 100 = 0.3. Exemplarily, based on the indoor path record data within a historical time period, a sample cumulative element set (i.e., the indoor element set of the user's cumulative path within a preset time range, such as the user's cumulative element set in the past hour {elevator, men's clothing store, bathroom}) and a sample path element set (i.e., the indoor element set of the user's path from the real-time location to the target location, such as the indoor element set of the user's path from A to B {elevator, bathroom, restaurant}) are extracted, and then the probability of the sample cumulative indoor element set and the two elements in the sample indoor element set appearing in the same path at the same time are calculated, and the correlation is marked. For example, the probability that "elevator" in the sample cumulative element set and "restaurant" in the sample indoor element set appear in the same path at the same time is 60%, then the correlation between "elevator" and "restaurant" is marked as 0.6. In this way, this step is repeated, and the indoor element correlation is marked according to the sample cumulative indoor element set and the sample indoor element set to obtain the sample indoor element correlation set.

[0062] Secondly, the model is constructed. Specifically, based on machine learning, a feature association analyzer is constructed. For example, the feature association analyzer is mainly composed of an input layer, a feature extraction layer, a core model layer, and an output layer. The input layer receives the encoding information of the accumulated indoor elements and the sample indoor elements (such as using one-hot encoding or word embedding vectors to represent elements such as elevators and coffee shops), and performs preprocessing on the data such as data cleaning (filtering invalid elements) and feature standardization (normalizing numerical features); the feature extraction layer uses an embedding layer to map the elements into low-dimensional semantic vectors (such as capturing the semantic association of elements), and combines the graph neural network (GNN) to extract the topological relationship features of the elements in the road network (such as the connectivity distance between elements). ); the core model layer can be based on the Bayesian network (based on the historical co-occurrence probability, modeling the conditional probability distribution between elements, and directly outputting the probability value of joint occurrence) or the multi-layer perceptron (MLP, suitable for scenarios where element features are represented by vectors, fitting the mapping relationship between element pairs and correlation through the fully connected layer, with a simple structure and high training efficiency.) architecture. Taking MLP as an example, the joint feature vector of the element pair is nonlinearly transformed through the fully connected layer and the ReLU activation function, and finally the Sigmoid function outputs a correlation value of 0-1 to represent the probability of the co-occurrence of the element pair; the output layer outputs the correlation.

[0063] Finally, the model is trained. Specifically, the sample cumulative indoor element set, the sample indoor element set and the sample indoor element correlation set are used to supervise the element correlation analyzer until convergence. Exemplarily, the sample cumulative indoor element set, the sample indoor element set and the sample indoor element correlation set are divided into a training set, a validation set and a test set according to 7:1.5:1.5, and indoor elements are traversed and selected in the sample indoor element set, and the model is input into the model with the indoor element combination in the sample cumulative indoor element set, and the element correlation is output. The labeled sample indoor element correlation is used as supervision, and the mean square error (MSE) or cross entropy loss function is used to compare the difference between the model predicted correlation and the labeled value, drive parameter optimization, introduce regularization terms (L1 / L2 regularization) to avoid overfitting, improve the generalization ability of the model, integrate optimization algorithms such as SGD and Adam, and update the model parameters through back propagation until the accuracy on the test set is greater than 95%, which is considered to be model convergence, stop training, and obtain the element correlation analyzer.

[0064] In summary, compared to existing technologies, this application analyzes the correlation between multiple indoor element sets and the accumulated indoor element correlations to obtain multiple element correlations. By calculating the correlation between the indoor elements of candidate paths and the user's accumulated indoor elements, the probability that different planned paths meet the user's current path selection preferences is quantified. This converts the user's historical behavioral preferences into a quantifiable path matching metric, making path planning more consistent with the user's personal preferences and providing the necessary data foundation for final path planning.

[0065] S40: Obtain multiple congestion parameters for multiple indoor paths, configure congestion weights and factor weights based on the user's historical movement path, combine the correlations of multiple factors, calculate multiple path scores, and screen to obtain an optimized indoor path as a path planning result.

[0066] The congestion level of a route and the types of indoor elements within the route (such as clothing stores, restrooms, elevators, etc.) are key considerations for users when choosing a route. Therefore, multiple routes can be comprehensively scored by integrating congestion parameters and element correlations. Furthermore, due to individual differences, different users have significant differences in their preferences for congestion levels and indoor elements when choosing a route. For example, some users prefer to avoid congested areas, while others may be more concerned about whether the selected route passes through the target store. Therefore, by analyzing the user's historical mobility data, their preference for congestion level or the upper limit of acceptance can be dynamically determined. Based on this, the weight ratio of congestion parameters and element correlations can be adjusted, and ultimately the optimal route planning result can be screened through a weighted scoring mechanism.

[0067] To address the above issues, this application obtains multiple congestion parameters for multiple indoor paths, configures congestion weights and factor weights based on the user's historical movement paths, combines the correlation between multiple factors, calculates multiple path scores, and screens the optimized indoor paths as path planning results.

[0068] Specifically, step S40 in the method includes:

[0069] Monitor the number of people in multiple indoor paths and calculate multiple crowding parameters;

[0070] Obtaining a user's historical movement path, and calculating an average user congestion parameter within the user's historical movement path;

[0071] Calculating a ratio of the average user congestion parameter to a preset congestion parameter, multiplying the ratio by a preset congestion weight to obtain a congestion weight, and calculating factor weights;

[0072] According to the congestion weight and the element weight, multiple congestion parameters and multiple element associations are combined to calculate multiple path scores, and the optimized indoor path is screened and obtained as the path planning result.

[0073] In the embodiment of the present application, the number of people in multiple indoor paths is first monitored and obtained, and multiple crowding parameters are calculated. Specifically, the real-time number of people in multiple indoor paths can be counted by indoor cameras, WiFi probes, Bluetooth beacons and other devices at a frequency of 5-10 seconds, and the density of people per unit area is calculated as the crowding parameter. The crowding parameter can reflect the crowding of multiple indoor paths. The larger the crowding parameter, the more crowded the path. For example, the number of people in multiple indoor paths is monitored in real time by indoor cameras, and the density of people per unit area (such as 3 people / m 2 4 people / m 2 2 people / m 2 ), multiple congestion parameters in multiple indoor paths are obtained, which can reflect the congestion conditions of different paths.

[0074] Secondly, the user's historical movement path is obtained, and the average user congestion parameter in the user's historical movement path is calculated. For example, based on historical data, the user's historical movement path is obtained, and the average user congestion parameter in the user's historical path is calculated. For example, the average user congestion parameter of a certain user is calculated to be 3 people / m 2 ,The average user congestion parameter can reflect the user’s ,preference for paths with different congestion levels or the acceptable ,congestion level.

[0075] Next, the ratio of the average user congestion parameter to the preset congestion parameter is calculated and multiplied by the preset congestion weight to obtain the congestion weight, and the factor weight is calculated, where congestion weight = (average user congestion parameter / preset congestion parameter) * preset congestion weight, and factor weight = 1 - congestion weight. Specifically, the preset congestion parameter can be set according to the specific application scenario. For example, for a shopping mall with a large flow of people or a shopping mall during peak hours, the preset congestion parameter can be set to 3 people / m 2 For shopping malls with low traffic volume, the preset crowding parameter can be appropriately lowered, such as setting the preset crowding parameter to 2 people / m 2 The preset congestion weight is dynamically adjusted according to the user's concern for congestion and indoor elements. For example, if the user is more concerned about the congestion in the path, the preset congestion weight can be appropriately increased, such as set to 0.6. If the user is more concerned about the indoor elements in the path, the preset congestion weight can be appropriately reduced, such as set to 0.4. For example, the preset congestion parameter is 3 people / m 2 The preset congestion weight is 0.6, and the average user congestion parameter is 3 people / m 2 , then the congestion weight = (3 people / m 2 / 3 people / m 2)*0.6=0.6, factor weight=1-0.6=0.4. In this way, the preset congestion weight can be modified according to the user's average user congestion parameter to obtain a congestion weight that is more in line with the user's actual selection preference.

[0076] Finally, according to the congestion weight and factor weight, combined with multiple congestion parameters and multiple factor correlations, multiple path scores are calculated and obtained, and the optimized indoor path is screened and obtained as the path planning result. Specifically, according to the congestion weight and factor weight, multiple congestion parameters and multiple factor correlations are weightedly calculated to obtain multiple path scores, and the indoor path corresponding to the maximum value of the multiple path scores is screened as the optimized indoor path to obtain the path planning result. Furthermore, the path score comprehensively considers the congestion parameter and the factor correlation, and the path score = congestion weight * congestion parameter + factor weight * factor correlation. Multiple path scores are obtained by calculation, and the path with the highest path score is selected as the path planning result. For example, when the congestion weight is 0.6 and the factor weight is 0.4, path 1: congestion parameter = 4 people / m 2 , element correlation = 0.8, then path score = 0.6*4+0.4×0.8 = 2.72; Path 2: congestion parameter = 2 people / m 2 , element correlation = 0.6, then the path score = 0.6*2+0.4×0.6=1.44, the path score of path 2 is higher than that of path 1, so path 2 is selected as the final path planning result.

[0077] Furthermore, the “calculating multiple path scores based on the congestion weight and the element weight, combining multiple congestion parameters and multiple element associations, and screening and obtaining an optimized indoor path as a path planning result” includes:

[0078] Performing weighted calculations on multiple congestion parameters and multiple factor associations based on the congestion weights and factor weights to obtain multiple path scores;

[0079] An indoor path corresponding to the maximum value of the multiple path scores is selected as the optimized indoor path to obtain a path planning result.

[0080] In the embodiment of the present application, first, based on the congestion weight and the element weight, a weighted calculation is performed on multiple congestion parameters and multiple element correlations to obtain multiple path scores, wherein the path score = congestion weight * congestion parameter + element weight * element correlation. The path score comprehensively considers the congestion parameter and the element correlation, and integrates the weights of the congestion parameter and the element correlation to score the current planned path. The higher the path score, the better the current path and the more it meets the user's personalized needs. For example, when the congestion weight is 0.6 and the element weight is 0.4, a weighted calculation is performed on multiple congestion parameters and multiple element correlations to obtain multiple path scores. For example, path 1: congestion parameter = 4 people / m 2 , element correlation = 0.8, then path score = 0.6*4+0.4×0.8 = 2.72; Path 2: congestion parameter = 2 people / m 2 , element correlation = 0.6, then the path score = 0.6*2+0.4×0.6=1.44.

[0081] Secondly, the indoor path corresponding to the maximum value of the multiple path scores is selected as the optimized indoor path to obtain the path planning result. Specifically, the higher the path score, the better the current path and the more it meets the user's personalized needs. Therefore, based on the multiple calculated path scores, the indoor path corresponding to the maximum score is selected as the optimized indoor path to obtain the path planning result. For example, the path score of path 1 is 2.72, and the path score of path 2 is 1.44. The path score of path 2 is higher than that of path 1, indicating that path 2 is better and more meets the user's personalized needs. Therefore, path 2 is selected as the final path planning result.

[0082] In summary, compared to existing technologies, this application obtains multiple congestion parameters for multiple indoor paths, assigns congestion weights and factor weights based on the user's historical travel history, and, combining multiple factor correlations, calculates multiple path scores. This process then selects optimized indoor paths as path planning results. By analyzing the user's historical travel data, the weighted ratios of congestion parameters and factor correlations are dynamically adjusted. A weighted scoring mechanism is then used to comprehensively score different paths based on these factors, ultimately selecting the optimal path planning result for user recommendations.

[0083] In summary, the embodiments of the present application have at least the following technical effects:

[0084] Compared to existing technologies, this application first obtains the user's real-time location and target location in the indoor environment, as well as the user's cumulative path, and then extracts the cumulative indoor elements within the cumulative path. By obtaining the user's real-time location, target location, and historical path information, it can accurately capture the user's recent preferences for indoor elements (such as elevators, shops, and restrooms) during path selection. This upgrades path planning from simple spatial distance planning to intelligent matching based on the user's personalized needs, and provides key data support for subsequent path optimization.

[0085] Secondly, this application plans multiple indoor paths based on the real-time location and target location, and extracts indoor elements within the multiple indoor paths to obtain multiple indoor element sets. This provides reliable data support for subsequent element correlation analysis and dynamic path planning.

[0086] Thirdly, this application analyzes the correlation between multiple indoor element sets and the accumulated indoor element correlations to obtain multiple element correlations. By calculating the correlation between the indoor elements of candidate paths and the user's accumulated indoor elements, the probability that different planned paths meet the user's current path selection preferences is quantified. This converts the user's historical behavioral preferences into a quantifiable path matching metric, making path planning more consistent with the user's personal preferences and providing the necessary data foundation for final path planning.

[0087] Finally, the application obtains multiple congestion parameters for multiple indoor paths. Based on the user's historical travel history, it assigns congestion weights and factor weights. Combined with the correlations between the multiple factors, it calculates multiple path scores and selects the optimized indoor paths as the path planning results. In this way, by analyzing the user's historical travel data, the weighted ratio of congestion parameters and factor correlations is dynamically adjusted. A weighted scoring mechanism is then used to comprehensively score different paths based on the congestion parameters and factor correlations, ultimately selecting the optimal path planning result for user recommendation.

[0088] Through the above technical solution, this application fully considers the differences in users' preferences for indoor elements and congestion levels within the path when choosing a path indoors, mines personalized preference characteristics from users' historical data, dynamically adjusts the path planning strategy based on real-time environmental data, and screens and outputs the optimal path planning results to provide users with accurate personalized path planning services.

[0089] Example 2, as Figure 2 As shown, based on the same inventive concept as the indoor path dynamic planning method provided in Example 1, an embodiment of the present invention also provides an indoor path dynamic planning system, including:

[0090] The data acquisition module 11 is used to obtain the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extract the cumulative indoor elements in the cumulative path;

[0091] A path generation module 12 is configured to plan and obtain multiple indoor paths based on the real-time location and the target location, and extract indoor elements within the multiple indoor paths to obtain multiple indoor element sets;

[0092] A correlation analysis module 13 is configured to analyze the correlations between the plurality of indoor element sets and the accumulated indoor elements to obtain correlations between the plurality of elements;

[0093] The optimization output module 14 is used to obtain multiple congestion parameters of multiple indoor paths, configure congestion weights and factor weights based on the user's historical movement path, combine the correlation between multiple factors, calculate multiple path scores, and screen to obtain the optimized indoor path as the path planning result.

[0094] The data acquisition module 11 is specifically used for:

[0095] Obtain the real-time location and target location of indoor users;

[0096] Get the user's path within the preset time range in the past and obtain the cumulative path;

[0097] A preset number of indoor elements closest to the real-time position in the cumulative path are extracted to obtain cumulative indoor elements.

[0098] The path generation module 12 is specifically configured to:

[0099] Obtain the road network of the indoor environment;

[0100] According to the road network, based on the real-time location and the target location, planning and obtaining multiple indoor paths with the shortest path lengths;

[0101] Indoor elements in the multiple indoor paths are extracted to obtain multiple indoor element sets of the multiple indoor paths.

[0102] The correlation analysis module 13 is specifically configured to:

[0103] Traversing and selecting indoor elements from multiple indoor element sets, combining the indoor elements with the accumulated indoor elements and inputting them into an element association analyzer, and outputting multiple indoor element association degree sets;

[0104] The mean of the correlation degree sets of multiple indoor elements is calculated to obtain the correlation degrees of multiple elements.

[0105] Furthermore, the training steps of the “element association analyzer” include:

[0106] Based on the indoor path record data in the historical time, the sample cumulative indoor element set and the sample indoor element set are collected, and the indoor element correlation degree of each sample cumulative indoor element and the sample indoor element is marked to obtain the sample indoor element correlation degree set, where the sample indoor element correlation degree includes the probability that the sample cumulative indoor element and the sample indoor element appear in the same indoor path at the same time;

[0107] Build a feature association analyzer based on machine learning;

[0108] The sample accumulated indoor element set, the sample indoor element set and the sample indoor element association degree set are used to perform supervised training on the element association analyzer until convergence.

[0109] The optimization output module 14 is specifically configured to:

[0110] Monitor the number of people in multiple indoor paths and calculate multiple crowding parameters;

[0111] Obtaining a user's historical movement path, and calculating an average user congestion parameter within the user's historical movement path;

[0112] Calculating a ratio of the average user congestion parameter to a preset congestion parameter, multiplying the ratio by a preset congestion weight to obtain a congestion weight, and calculating factor weights;

[0113] According to the congestion weight and the element weight, multiple congestion parameters and multiple element associations are combined to calculate multiple path scores, and the optimized indoor path is screened and obtained as the path planning result.

[0114] Furthermore, the “calculating multiple path scores based on the congestion weight and the element weight, combining multiple congestion parameters and multiple element associations, and screening and obtaining an optimized indoor path as a path planning result” includes:

[0115] Performing weighted calculations on multiple congestion parameters and multiple factor associations based on the congestion weights and factor weights to obtain multiple path scores;

[0116] An indoor path corresponding to the maximum value of the multiple path scores is selected as the optimized indoor path to obtain a path planning result.

[0117] In summary, the embodiments of the present application have at least the following technical effects:

[0118] Compared with the existing technology, this application first obtains the real-time location and target location of indoor users, as well as the user's cumulative path, through the data acquisition module, and extracts the cumulative indoor elements in the cumulative path. By obtaining the user's real-time location, target location and historical path information, it can accurately capture the user's recent preference for indoor elements (such as elevators, shops, toilets, etc.) in path selection, so that path planning is upgraded from simple spatial distance planning to intelligent matching based on user personalized needs, and provides key data support for subsequent path optimization. Secondly, through the path generation module, based on the real-time location and target location, multiple indoor paths are planned and obtained, and indoor elements in multiple indoor paths are extracted to obtain multiple indoor element sets, which provides reliable data support for subsequent element correlation analysis and path dynamic planning. Next, the correlation analysis module analyzes the correlations between multiple indoor element sets and the accumulated indoor element values ​​to obtain multiple element correlations. By calculating the correlations between the indoor elements of candidate paths and the user's accumulated indoor elements, the probability that different planned paths meet the user's current path selection preferences is quantified. This transforms the user's historical behavioral preferences into quantifiable path matching metrics, making path planning more consistent with the user's personal preferences and providing the necessary data foundation for final path planning. Finally, the optimization output module obtains multiple congestion parameters for multiple indoor paths. Based on the user's historical movement paths, congestion weights and element weights are assigned. Combined with the multiple element correlations, multiple path scores are calculated and selected. The optimized indoor paths are then selected as the path planning results. By analyzing the user's historical movement data, the weights of the congestion parameters and element correlations are dynamically adjusted. A weighted scoring mechanism is then used to comprehensively score different paths based on the congestion parameters and element correlations. Ultimately, the optimal path planning result is selected and recommended to the user. This allows for personalized path planning services.

[0119] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0120] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0121] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0122] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0123] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0124] Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts become known.

[0125] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. Indoor path dynamic planning method, characterized by: The method comprises: Obtaining the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extracting the cumulative indoor elements within the cumulative path; Planning and obtaining a plurality of indoor paths according to the real-time position and the target position, and extracting indoor elements within the plurality of indoor paths to obtain a plurality of indoor element sets; Analyzing the correlation between the plurality of indoor element sets and the accumulated indoor element to obtain a plurality of element correlations; Multiple congestion parameters of multiple indoor paths are obtained. Based on the user's historical movement path, congestion weights and factor weights are configured. Combined with the correlation between multiple factors, multiple path scores are calculated and selected to obtain the optimized indoor path as the path planning result.

2. The indoor path dynamic planning method according to claim 1, characterized in that: Obtaining the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extracting the cumulative indoor elements within the cumulative path, including: Obtain the real-time location and target location of indoor users; Get the user's path within the preset time range in the past and obtain the cumulative path; A preset number of indoor elements closest to the real-time position in the cumulative path are extracted to obtain cumulative indoor elements.

3. The indoor path dynamic planning method according to claim 1, characterized in that: Planning and obtaining multiple indoor paths based on the real-time location and the target location, and extracting indoor elements within the multiple indoor paths to obtain multiple indoor element sets, including: Obtain the road network of the indoor environment; According to the road network, based on the real-time location and the target location, planning and obtaining multiple indoor paths with the shortest path lengths; Indoor elements in the multiple indoor paths are extracted to obtain multiple indoor element sets of the multiple indoor paths.

4. The indoor path dynamic planning method according to claim 1, characterized in that: Analyzing the associations of the plurality of indoor element sets and the accumulated indoor element to obtain the associations of the plurality of elements includes: Traversing and selecting indoor elements from multiple indoor element sets, combining the indoor elements with the accumulated indoor elements and inputting them into an element association analyzer, and outputting multiple indoor element association degree sets; The mean of the correlation degree sets of multiple indoor elements is calculated to obtain the correlation degrees of multiple elements.

5. The indoor path dynamic planning method according to claim 4, characterized in that: The training steps of the feature association analyzer include: Based on the indoor path record data in the historical time, the sample cumulative indoor element set and the sample indoor element set are collected, and the indoor element correlation degree of each sample cumulative indoor element and the sample indoor element is marked to obtain the sample indoor element correlation degree set, where the sample indoor element correlation degree includes the probability that the sample cumulative indoor element and the sample indoor element appear in the same indoor path at the same time; Build a feature association analyzer based on machine learning; The sample accumulated indoor element set, the sample indoor element set and the sample indoor element association degree set are used to perform supervised training on the element association analyzer until convergence.

6. The indoor path dynamic planning method according to claim 1, characterized in that: Obtain multiple congestion parameters for multiple indoor paths, configure congestion weights and factor weights based on the user's historical movement path, calculate multiple path scores based on the correlation between multiple factors, and screen and obtain the optimized indoor path as the path planning result, including: Monitor the number of people in multiple indoor paths and calculate multiple crowding parameters; Obtaining a user's historical movement path, and calculating an average user congestion parameter within the user's historical movement path; Calculating a ratio of the average user congestion parameter to a preset congestion parameter, multiplying the ratio by a preset congestion weight to obtain a congestion weight, and calculating factor weights; According to the congestion weight and the element weight, multiple congestion parameters and multiple element associations are combined to calculate multiple path scores, and the optimized indoor path is screened and obtained as the path planning result.

7. The indoor path dynamic planning method according to claim 6, characterized in that: Based on the congestion weight and the element weight, multiple congestion parameters and multiple element associations are combined to calculate multiple path scores, and the optimized indoor path is screened and obtained as the path planning result, including: According to the congestion weight and the factor weight, a weighted calculation is performed on multiple congestion parameters and multiple factor associations to obtain multiple path scores; An indoor path corresponding to the maximum value of the multiple path scores is selected as the optimized indoor path to obtain a path planning result.

8. Indoor path dynamic planning system, characterized by: Used to perform the method according to any one of claims 1 to 7, comprising: A data acquisition module is used to obtain the real-time location and target location of the user in the indoor environment, as well as the user's cumulative path, and extract the cumulative indoor elements within the cumulative path; A path generation module is used to plan and obtain multiple indoor paths based on the real-time position and the target position, and extract indoor elements in the multiple indoor paths to obtain multiple indoor element sets; a correlation analysis module, configured to analyze the correlations between the plurality of indoor element sets and the accumulated indoor elements to obtain a plurality of element correlations; The optimization output module is used to obtain multiple congestion parameters of multiple indoor paths, configure congestion weights and factor weights based on the user's historical movement path, combine the correlation of multiple factors, calculate multiple path scores, and screen to obtain the optimized indoor path as the path planning result.

Images