Urban inland inundation emergency path planning method and system based on multi-modal crowdsourcing data and dynamic risk weight
By combining multimodal crowdsourced data with dynamic risk weights, the road network status is updated in real time, solving the problems of data lag and insufficient risk assessment in existing route planning systems during urban flooding disasters, and realizing safe, reliable and adaptive evacuation route planning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-30
- Publication Date
- 2026-04-14
AI Technical Summary
Existing route planning systems suffer from unreliable and unsafe evacuation routes when faced with sudden disasters such as urban flooding due to data lag, static risk models, and a lack of intelligent integration with crowdsourced data.
By combining multimodal crowdsourced data with dynamic risk weights, the road network status is updated in real time. Using multimodal data reported by user terminals, data preprocessing, standardization, multi-source verification, and weight calculation are performed to dynamically assess road segment risks. Combined with static road attributes, the passage cost is updated to plan the optimal safe evacuation route.
It significantly improves evacuation safety, has strong real-time performance and robustness, and can quickly respond and automatically adjust path planning strategies to ensure the reliability and adaptability of paths.
Smart Images
Figure CN121855563A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent navigation and emergency management technology, and in particular to a method and system for safe path planning using multimodal crowdsourced real-time data and dynamic risk weighting models in the event of sudden natural disasters such as urban flooding. Background Technology
[0002] With rapid urbanization and frequent extreme weather events, urban flooding has become a major threat to public safety. In emergency evacuation scenarios, the public heavily relies on electronic maps for navigation. However, the core optimization goal of mainstream route planning services (such as Gaode Maps and Baidu Maps) is traffic efficiency (shortest time or distance), and their technological foundation depends on historical traffic flow, static road networks, and real-time congestion data. These systems have fundamental shortcomings when facing disasters like urban flooding where road conditions change rapidly, specifically as follows: (1) Data lag and insufficient coverage. The existing system relies on fixed sensors deployed by the authorities, which results in slow data updates, many blind spots, and an inability to quickly respond to sudden water accumulation points, making it difficult to form a real-time risk situation map with full coverage.
[0003] (2) Static risk model. The path cost weights (such as congestion weights) of traditional algorithms are static or semi-static, which cannot characterize the dynamic changes in risks such as water depth and flow velocity during urban flooding, and cannot achieve the planning goal of "safety first".
[0004] (3) Lack of intelligent integration of crowdsourced data. Although some applications have attempted to introduce user reports, there is a lack of effective processing mechanisms for the inherent characteristics of crowdsourced data (such as heterogeneity, high noise, and inconsistent credibility), making it impossible to conduct credibility assessment and multi-source verification, resulting in unreliable planning results.
[0005] (4) Mismatched application scenarios. In existing emergency evacuation plans (such as those based on sound analysis or building elevator scheduling), the data source is weakly correlated with the risk of road flooding; or the scenarios are specific and cannot be applied to urban-level, open road public flood evacuation. Summary of the Invention
[0006] To address at least some of the aforementioned problems, this invention proposes a method and system for urban flooding emergency route planning based on multimodal crowdsourced data and dynamic risk weights.
[0007] In a first aspect, the present invention provides an urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights, including: a road network status update process and a user route search process; The road network status update process includes: Step A1, Data Reporting: The user terminal reports multimodal data reflecting the depth of water accumulation; Step A2, Data Reception and Preprocessing: The application server receives data packets from the user terminal and performs format verification and cleaning on the data packets; Step A3, Data Type Judgment and Branch Processing: Determine the data type of the multimodal data in order to use the corresponding processing method to obtain the water depth; Step A4, Data Standardization and Multi-Source Validation: Unify the multimodal data reported by different user terminals into a standard format, and perform cross-validation on the relevant multimodal data within a spatial range; Step A5, Data Fusion Processing and Weight Calculation: Considering the timeliness and spatial consistency of the data, calculate a dynamic comprehensive credibility weight for each multimodal data point; Step A6, Multimodal Data Fusion and Road Segment Risk Assessment: For each road segment in the road network, aggregate its relevant multimodal data, and calculate the comprehensive assessment water depth of the road segment based on the comprehensive credibility weight calculated in Step A5, thereby determining the water depth weight factor of the road segment. Step A7, periodically execute the spatiotemporal decay model: periodically calculate the comprehensive credibility weight of historical data based on the spatiotemporal decay model, realize data aging processing, and update the global road network risk map; The user path search process includes: Step B1, User enters destination: Set the evacuation target point on the user terminal and initiate a request; Step B2, Actively apply the spatiotemporal decay model: When a user makes a request, obtain the latest global road network risk map after processing by the spatiotemporal decay model; Step B3, Update the dynamic risk cost function: Based on the latest global road network risk map, refresh the passage cost of each road segment in the global road network risk map by combining the static attributes of the road and the water depth weight factor of the road segment; Step B4, Search for the optimal path: Run the path planning algorithm to search for the optimal safe evacuation path from the starting point to the destination with the goal of minimizing the travel cost; Step B5, Output Path: Send the found optimal safe evacuation path to the user terminal for navigation guidance.
[0008] Furthermore, the data type of the multimodal data is determined in order to use the corresponding processing method to obtain the water depth, specifically including: When the water depth reflected by the multimodal data is the water depth measured by the user through the tool, the water depth is read directly. When the water depth reflected by the multimodal data is a descriptive body part and its height, the body part-height ratio mapping table is looked up to obtain the ratio coefficient between the body part and the height, and the water depth is calculated based on the ratio coefficient and the height.
[0009] Furthermore, considering the timeliness and spatial consistency of the data, a dynamic comprehensive credibility weight is calculated for each piece of multimodal data. Specifically, this includes calculating the comprehensive credibility weight of the i-th piece of multimodal data according to the following formula. ; in, ; As a spatiotemporal decay weight used to characterize the timeliness of data, For multi-source verification weights used to characterize the consistency of the data space, The attenuation coefficient is... Let N be the difference between the reporting time of the i-th multimodal data point and the current time, and let N represent the number of neighboring points of the water accumulation point corresponding to the i-th multimodal data point. Z-score is used to characterize the difference between a given data point and the mean of its N neighboring data points. This is the preset Z-score threshold.
[0010] Furthermore, step A6 specifically includes: Road segment waterlogging data aggregation: For road segments Collect all reported locations located on or directly affected by this road section, forming a set of water accumulation information for that road section. ; Calculating the comprehensive assessment of water accumulation depth for road sections: A weighted average method is used, incorporating aggregate data. Water depth in all data and its overall credibility weight Calculate the road segment Comprehensive assessment of water depth : in, For the first The water depth of each data point; Determine the water depth weighting factor for road sections: Multiple water depth intervals are pre-set, and a water depth weighting factor is assigned to each interval, with a larger weighting factor for deeper water. This allows for the determination of the comprehensive assessment of water accumulation depth for the road section. The water depth weighting factor of the road section is determined by the water depth range it falls within. .
[0011] Furthermore, the toll cost of each road segment in the global road network risk map is updated by combining the static attributes of the road and the water depth weight factor of the road segment. Specifically, this includes calculating the toll cost of road segment j according to the following formula. ; in, Representative road section Length, Representative road section Travel time per unit length under ideal conditions Indicates road segment The cost of passage under ideal conditions. Indicates road segment The water depth weighting factor is positively correlated with the water depth. When the water depth exceeds a threshold, Set to infinity make the road section The cost is infinitely high and therefore avoided; Indicates road segment The road width weighting factor is positively correlated with the road width; Indicates road segment The altitude weighting factor is positively correlated with altitude.
[0012] Furthermore, the path planning algorithm, with the goal of minimizing travel cost, searches for the optimal safe evacuation route from the starting point to the destination, specifically including: Starting from the user terminal's current location and ending at the designated evacuation target point, the calculated travel cost is used as the starting point. Using the weights of the edges corresponding to road segment j, the graph search algorithm is run to search for the path with the lowest travel cost in the global road network risk map, which is then used as the optimal safe evacuation path.
[0013] Secondly, this invention provides an urban flooding emergency route planning system based on multimodal crowdsourced data and dynamic risk weights, comprising: The user terminal is used to report multimodal data reflecting the depth of water accumulation; to allow users to set evacuation target points and initiate requests; and to receive the optimal safe evacuation route for navigation guidance. An application server receives data packets from user terminals and performs format verification and cleaning on the data packets; determines the data type of multimodal data to obtain the water depth using corresponding processing methods; unifies the multimodal data reported by different user terminals into a standard format and performs cross-validation on related multimodal data within a spatial range; calculates a dynamic comprehensive credibility weight for each multimodal data piece based on the timeliness and spatial consistency of the data; aggregates related multimodal data for each road segment in the road network and calculates the comprehensive assessment water depth of the road segment based on the comprehensive credibility weight, thereby determining the water depth weight factor of the road segment; periodically calculates the comprehensive credibility weight of historical data based on a spatiotemporal decay model to achieve data aging processing and update the global road network risk map; when a user initiates a request, obtains the latest global road network risk map after processing by the spatiotemporal decay model; based on the latest global road network risk map, and combined with the static attributes of the road and the water depth weight factor of the road segment, refreshes the passage cost of each road segment in the global road network risk map; A spatial database is used to store multimodal data from user terminals and dynamic global road network risk maps; The path planning engine runs path planning algorithms to search for the optimal safe evacuation route from the starting point to the destination with the goal of minimizing travel costs, and then sends the found optimal safe evacuation route to the user terminal.
[0014] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as described in the first aspect.
[0015] Fourthly, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.
[0016] The beneficial effects of this invention are as follows: (1) Significantly improves evacuation safety. By calculating the passage cost of each road segment, this invention fundamentally avoids high-risk road sections such as deep water accumulation areas, thus ensuring the safety of users' lives and property.
[0017] (2) It has extremely strong real-time performance. This invention utilizes the power of the masses to quickly perceive the traffic conditions of the entire network, and its response speed far exceeds that of fixed sensor networks.
[0018] (3) High robustness. This invention effectively filters out erroneous and malicious data by using an intelligent data fusion model based on spatiotemporal decay and multi-source verification, and retains reliable multimodal data to generate a risk map, thereby ensuring the reliability of the system's output path under information noise.
[0019] (4) Good adaptability. The risk map maintained by this invention can be automatically adjusted according to the disaster situation (deepening, receding or shifting of water). In this way, when route planning is carried out based on the adaptively adjusted risk map, the rationality of the planning strategy can always be maintained. Attached Figure Description
[0020] Figure 1 A flowchart of an urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights is provided for an embodiment of the present invention; Figure 2 This is a schematic diagram of spatiotemporal decay weights provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of multi-source verification weights provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of path planning in traditional methods that do not consider safety costs. Figure 5 A schematic diagram of path planning considering safety costs provided for embodiments of the present invention. Figure 6 A schematic diagram of the structure of an urban flooding emergency route planning system based on multimodal crowdsourced data and dynamic risk weights provided in an embodiment of the present invention; Figure 7 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation
[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0022] This invention aims to construct a path planning scheme that integrates data perception, fusion processing, and intelligent decision-making, transforming unreliable crowdsourced information into highly credible dynamic risk information, and thereby driving safe path planning.
[0023] like Figure 1 As shown, this embodiment of the invention provides an urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights. This method is a dual-process asynchronous collaboration mode, thereby decoupling the system's data maintenance function from its real-time service function.
[0024] Specifically, the road network status update process is a continuously running background task responsible for receiving and processing all crowdsourced data reported by users, maintaining a global, real-time updated road risk knowledge base. It does not rely on requests from individual users but serves all users, such as... Figure 1 The blue flowchart on the left shows the user route search process, a real-time task triggered by a user request. When a user needs route planning, it retrieves the latest road network status from the knowledge base built by the background task and quickly completes personalized route planning, such as... Figure 1 The green flow chart on the right side shows the two flows. The two flows interact with each other through a "spatiotemporal decay model" to ensure the real-time nature and accuracy of the data used for path planning.
[0025] The road network status update process is continuously responsible for sensing and updating the real-time risk status of the entire road network. Its execution steps are as follows: (1) Data reporting: Users report multimodal data reflecting water depth through the user terminal App; the data can be numerical measurement values (hereinafter referred to as measured values) or descriptive information (hereinafter referred to as descriptive values).
[0026] (2) Data reception and preprocessing: The application server receives data packets from the user terminal and performs format verification and basic cleaning on the received data packets.
[0027] (3) Data type judgment and branch processing: The application server judges the reported data type and enters the corresponding processing branch: (a) Obtain water depth directly: If it is a measured value, then read its value directly.
[0028] (b) Calculate water depth by looking up the mapping table: If it is a descriptive value (such as "water depth up to the knee"), the specific water depth is calculated by combining the user's height with a predefined mapping table.
[0029] Specifically, the measured value refers to the water depth (in centimeters) measured by the user using the tool. The descriptive value refers to the descriptive body part (such as "knee") selected by the user and their height.
[0030] For the described values, embodiments of the present invention have designed a key body part-height ratio mapping table, as shown in Table 1. This table is intelligently converted into a water depth with uniform dimensions, solving the problem that non-professional users cannot accurately measure the depth.
[0031] Table 1. Examples of Body Part-Height Ratio Mapping (4) Data standardization and multi-source verification: unify the multimodal data reported by different user terminals into a standard format, and cross-verify the relevant multimodal data within the spatial range (such as the consistency judgment of multiple points within a 50-meter radius).
[0032] (5) Data fusion processing and weight calculation: Based on the timeliness and spatial consistency of the data, a dynamic comprehensive credibility weight is calculated for each data point. The timeliness of the data is obtained through a spatiotemporal decay model; the spatial consistency of the data is obtained through multi-source verification in the previous step.
[0033] (6) Multimodal data fusion and road segment risk assessment: For each road segment in the road network, aggregate its related multimodal data, and calculate the comprehensive assessment water depth of the road segment based on the comprehensive credibility weight calculated in step (5), and then determine the water depth weight factor of the road segment. (7) Periodically execute the spatiotemporal decay model: This step is the loop control point of the process. The system periodically (e.g., every 5 minutes) recalculates the comprehensive credibility weight of historical data based on the spatiotemporal decay model to realize data aging and update the global road network risk map. This dynamically updated global road network risk map forms the decision basis for the real-time path search on the right.
[0034] As one possible implementation, when new crowdsourced data is reported, a re-evaluation of historical data of spatially adjacent points is triggered. If the overall credibility weight of the new data is high and inconsistent with the historical data, the weight of the historical data is penalized by lowering it.
[0035] The user route search process is the service interface through which the system interacts with the user, providing real-time and secure route planning services. Its steps are as follows: (1) User inputs destination: The user sets the evacuation target point on the user interface through the user terminal and initiates a request.
[0036] (2) Actively apply the spatiotemporal decay model: This is a key step connecting the left and right processes. When a user makes a request, this process will actively obtain the latest and most accurate global road network risk map after processing by the spatiotemporal decay model from the global state maintained by the left process.
[0037] (3) Update the dynamic risk cost function: Based on the latest global road network risk map, combined with the static attributes of the road (such as length, width and altitude) and the water depth weight factor of the road segment, quickly refresh the passage cost (also known as safety cost) of each road segment in the global road network risk map.
[0038] (4) Search for the optimal path: Run a path planning algorithm (such as the A* algorithm) to search for the optimal safe evacuation path from the starting point to the end point with the goal of "lowest travel cost".
[0039] (5) Output path: The planned optimal safe evacuation path is sent to the user terminal for navigation guidance.
[0040] The urban flooding emergency route planning method provided in this embodiment of the invention utilizes the power of the masses to quickly perceive the road conditions of the entire network, and the response speed is far greater than that of fixed sensor networks, thus having extremely strong real-time performance; through the intelligent data fusion model in steps (4) to (6) of the road network status update process, it effectively filters out erroneous and malicious data, ensuring the reliability of the system's output under information noise, thus having high robustness; dynamically updates the global road network risk map, so that it can automatically adjust with the disaster situation (deepening, receding or shifting of water), and always maintain the rationality of the planning strategy, thus having good adaptability.
[0041] In one embodiment, step (5) of the road network status update process specifically includes: The system calculates a dynamic comprehensive confidence weight for each piece of multimodal data. This weight consists of two parts: the spatiotemporal decay weight. and multi-source verification weights The overall credibility weight of the i-th multimodal data point It can be represented as: (1) in .
[0042] In addition, it should be noted that before calculating the overall credibility weight, the system will perform a preliminary usability check on the data and remove obviously invalid data (such as data with format errors).
[0043] Spatiotemporal decay weights Based on the difference between the reported time and the current time The calculation shows an exponential decay over time, as shown in the following formula: (2) in This is the attenuation coefficient, which can be adjusted based on local drainage conditions, rainfall intensity, soil permeability, etc. For areas with different attenuation coefficients, Figure 2 The calculation results of the spatiotemporal attenuation weights for the two regions are given.
[0044] Multi-source validation weights are used to measure the spatial consistency of data. Their core function is to identify outliers and unreliable data through comparison with neighboring data. The specific calculation process is as follows: First, a reasonable search radius is set, centered on the water accumulation point corresponding to the current multimodal data. Next, the system will look up the radius. The number of neighboring points is denoted as the number of all valid neighboring points reported (whose data has undergone preliminary availability checks). Then, calculate the water depth at the current point and the average water depth at all neighboring points. The difference, and calculate the difference relative to the standard deviation of water depth at neighboring points. The multiple of (i.e., Z-score) is calculated using the formula: .like Within acceptable limits (e.g.) 2. If the difference between the current point's water depth and the average value of neighboring points is within 2 standard deviations, then the spatial consistency is considered high; ultimately, the weights... The weighting is determined by both the number of neighboring points and the calculated consistency index, ensuring that a higher weight is given when there are a sufficient number of neighboring reference points and high spatial consistency. The calculation formula is as follows: (3) in, To preset the Z-score threshold (e.g., T=2.5), when At that time, the second part will make This significantly reduces the number of outliers in the space, thus effectively penalizing them.
[0045] A diagram illustrating the weights of multi-source validation is shown below. Figure 3 As shown.
[0046] The overall credibility weight of each multimodal data point is calculated. Next, the system performs a screening of trusted data: With a preset confidence threshold Comparison. Only when... Only when the threshold is met will the data be retained and marked as "trustworthy multimodal data". For data that fails the threshold comparison, the system can determine its appropriate classification based on its... Values are categorized (e.g., marked as "to be verified" or "unavailable") and can be selectively logged for later analysis to optimize model parameters. This filtering process effectively removes erroneous data resulting from non-manual input errors and malicious injections.
[0047] In one embodiment, step (6) of the above embodiment fuses multimodal data to assess the overall water accumulation situation of the road section, as follows: Road segment waterlogging data aggregation: For road segments Collect all reported multimodal data located on or directly affecting this road segment, starting with the water accumulation information set for that road segment. .
[0048] Calculating the comprehensive assessment of water accumulation depth for road sections: A weighted average method is used, incorporating aggregate data. Water depth in all data and its credibility weight Calculate the road segment Comprehensive assessment of water depth : (4) in, The first one calculated according to the aforementioned method The water depth of each data point This is a collection of valid water accumulation information for this road section.
[0049] Determine the water depth weighting factor for road sections: Multiple water depth intervals are pre-set, and a water depth weighting factor is assigned to each interval, with a larger weighting factor for deeper water. This allows for the determination of the comprehensive assessment of water accumulation depth for the road section. The water depth weighting factor of the road section is determined by the water depth range it falls within. The depth unit is set to centimeters, and the upper and lower thresholds for each water depth range can be adjusted according to actual conditions. For example, when... When it is 50cm, set it as This causes the section of road to be avoided in route planning.
[0050] One possible implementation is as follows: (5) In one embodiment, updating the dynamic risk cost function in the user route search process specifically includes: abstracting the urban road network into a graph structure, and calculating the dynamic travel cost for each road segment based on the processed reliable multimodal data. : (6) in, Representative road section Length; Representative road section In ideal conditions (e.g., no water accumulation, no congestion), the travel time per unit length (minutes / km) or the generalized cost, Indicates road segment The cost of passage under ideal conditions; For road section The water depth weighting factor can be determined based on the water depth in or near the road section. (After confidence-weighted averaging) the mapping is obtained; as the water depth increases, The larger the value, the more likely it is that when the water depth exceeds the threshold (e.g., 50cm), Set to infinity make the road section The cost is infinitely high and therefore it is avoided. This is a road width weighting factor; the wider the road, the higher the traffic safety. The larger the value. Altitude is a weighted factor; the lower the altitude of a road segment, the higher the potential risk. The smaller the value.
[0051] In one embodiment, step (4) in the user route search process specifically includes: taking the current location of the user terminal as the starting point and the set evacuation target point as the ending point, and using the calculated dynamic passage cost... As an edge (i.e., road segment) The weights of the corresponding edges in the risk map are used to run a graph search algorithm (such as A* algorithm or Dijkstra algorithm) to find the path with the lowest overall travel cost, which is the optimal safe evacuation path.
[0052] Taking a region affected by flooding as an example, the method of the present invention is carried out according to the following process: System initialization: Import urban road network data and build a graph model.
[0053] Data Report: User A reported "latitude and longitude (116.397128, 39.916527), water depth description 'knee', height 180cm". The system automatically converted it to a water depth of 49cm.
[0054] Data fusion: Five minutes later, user B reported a measured water depth of 55cm from nearby. The system calculated the spatiotemporal attenuation weights and multi-source verification weights of both data, arriving at a high-confidence result: the weighted average water depth for this road segment is approximately 52cm.
[0055] Route planning: The system calculates the cost of each route segment, including the segment with 52cm of standing water, which is avoided due to its extremely high risk. The system plans a safe route around the risky area for the user requesting navigation and sends the route to the terminal for voice and graphic navigation.
[0056] Figure 4 The traditional path calculation results without considering road safety are presented. Figure 5 The path calculation results for dynamic travel costs considering safety, as proposed in this invention, are presented. Figure 4 and Figure 5 In the diagram, the colors of the square areas represent different depths of water. It can be seen that when traditional methods do not consider safe travel, the paths planned by these methods directly traverse dangerous areas. However, this invention, by considering safety and cost, plans paths that bypass dangerous areas.
[0057] This invention proposes an automatic conversion and fusion mechanism for multimodal crowdsourced data. In particular, it quantifies qualitative descriptions through a "body part-height ratio mapping table," significantly lowering the data collection threshold and expanding data sources. A dynamic credibility weight model combining spatiotemporal decay and multi-source verification is created. This model automatically assesses data reliability and dynamically updates with time and new data, effectively solving the noise and credibility problems of crowdsourced data, outperforming simple averaging or voting mechanisms. A dynamic risk cost function using the road segment's passage cost as its safety factor is designed, transforming route planning from the traditional "efficiency optimization" paradigm to a "feasibility optimization under safety constraints" paradigm, thus enabling emergency route planning.
[0058] Based on the same inventive concept, such as Figure 6 As shown, this embodiment of the invention provides an urban flooding emergency route planning system based on multimodal crowdsourced data and dynamic risk weights, including a user terminal, an application server, a spatial database, and a route planning engine.
[0059] Specifically, the user terminal is used to report multimodal data reflecting water depth; to allow users to set evacuation target points and initiate requests; and to receive optimal safe evacuation routes for navigation guidance. The application server is used to receive data packets from the user terminal and perform format verification and cleaning on the data packets; to determine the data type of the multimodal data so as to use the corresponding processing method to obtain the water depth; to unify the multimodal data reported by different user terminals into a standard format and to perform cross-validation on related multimodal data within a spatial range; to calculate a dynamic comprehensive credibility weight for each multimodal data based on the timeliness and spatial consistency of the data; and to periodically adjust the data according to spatiotemporal decay. The model calculates the comprehensive credibility weight of historical data to achieve data aging and updates the global road network risk map. When a user makes a request, the latest global road network risk map, processed by the spatiotemporal decay model, is obtained. Based on the latest global road network risk map, the toll cost of each road segment in the global road network risk map is refreshed by combining the static attributes of the roads. The spatial database is used to store multimodal data from user terminals and dynamic global road network risk maps. The path planning engine is used to run path planning algorithms to search for the optimal safe evacuation route from the origin to the destination with the goal of minimizing toll costs, and then sends the searched optimal safe evacuation route to the user terminal.
[0060] It should be noted that the urban flooding emergency route planning system provided in this embodiment of the invention is for implementing the above methods. Its specific functions can be referred to in the above method embodiments, and will not be repeated here.
[0061] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7As shown, the electronic device may include: a processor 701, a communication interface 702, a memory 703, and a communication bus 704, wherein the processor 701, the communication interface 702, and the memory 703 communicate with each other through the communication bus 704. The processor 701 can call logical instructions in the memory 703 to execute an urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights. This method includes: a road network status update process and a user route search process. The road network status update process includes: Step A1, data reporting: user terminals report multimodal data reflecting water depth; Step A2, data receiving and preprocessing: the application server receives data packets from user terminals and performs format verification and cleaning on the data packets; Step A3, data type judgment and branch processing: the data type of the multimodal data is determined so that the corresponding processing method can be used to obtain the water depth; Step A4, data standardization and multi-source verification: the multimodal data reported by different user terminals is unified into a standard format, and related multimodal data is cross-validated within a spatial range. Step A5, Data Fusion Processing and Weight Calculation: Considering the timeliness and spatial consistency of the data, calculate a dynamic comprehensive credibility weight for each multimodal data point; Step A6, Multimodal Data Fusion and Road Segment Risk Assessment: For each road segment in the road network, aggregate its relevant multimodal data, and based on the comprehensive credibility weight calculated in Step A5, calculate the comprehensive assessment water depth of the road segment, thereby determining its water depth weight factor; Step A7, Periodic Execution of the Spatiotemporal Decay Model: Periodically calculate the comprehensive credibility weight of historical data based on the spatiotemporal decay model to achieve data aging processing and update the global road network risk map; the user road The path search process includes: Step B1, User inputs destination: The user sets the evacuation target point on the user terminal and initiates a request; Step B2, Actively applies the spatiotemporal decay model: When the user initiates a request, the latest global road network risk map processed by the spatiotemporal decay model is obtained; Step B3, Updates the dynamic risk cost function: Based on the latest global road network risk map, the passage cost of each road segment in the global road network risk map is refreshed by combining the static attributes of the road and the water depth weight factor of the road segment; Step B4, Searches for the optimal path: The path planning algorithm is run to search for the optimal safe evacuation path from the starting point to the destination with the goal of minimizing the passage cost; Step B5, Outputs the path: The optimal safe evacuation path found is sent to the user terminal for navigation guidance.
[0062] Furthermore, when the logical instructions in the aforementioned memory 703 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0063] This invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can execute the urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights provided in the above-described method embodiments.
[0064] This invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the urban flooding emergency path planning method based on multimodal crowdsourced data and dynamic risk weights provided in the above-described method embodiments.
[0065] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for urban flooding emergency route planning based on multimodal crowdsourced data and dynamic risk weights, characterized in that, include: Road network status update process and user route search process; The road network status update process includes: Step A1, Data Reporting: The user terminal reports multimodal data reflecting the depth of water accumulation; Step A2, Data Reception and Preprocessing: The application server receives data packets from the user terminal and performs format verification and cleaning on the data packets; Step A3, Data Type Judgment and Branch Processing: Determine the data type of the multimodal data in order to use the corresponding processing method to obtain the water depth; Step A4, Data Standardization and Multi-Source Validation: Unify the multimodal data reported by different user terminals into a standard format, and perform cross-validation on the relevant multimodal data within a spatial range; Step A5, Data Fusion Processing and Weight Calculation: Considering the timeliness and spatial consistency of the data, calculate a dynamic comprehensive credibility weight for each multimodal data point; Step A6, Multimodal Data Fusion and Road Segment Risk Assessment: For each road segment in the road network, aggregate its relevant multimodal data, and calculate the comprehensive assessment water depth of the road segment based on the comprehensive credibility weight calculated in Step A5, thereby determining the water depth weight factor of the road segment. Step A7, periodically execute the spatiotemporal decay model: periodically calculate the comprehensive credibility weight of historical data based on the spatiotemporal decay model, realize data aging processing, and update the global road network risk map; The user path search process includes: Step B1, User enters destination: Set the evacuation target point on the user terminal and initiate a request; Step B2, Actively apply the spatiotemporal decay model: When a user makes a request, obtain the latest global road network risk map after processing by the spatiotemporal decay model; Step B3, Update the dynamic risk cost function: Based on the latest global road network risk map, refresh the passage cost of each road segment in the global road network risk map by combining the static attributes of the road and the water depth weight factor of the road segment; Step B4, Search for the optimal path: Run the path planning algorithm to search for the optimal safe evacuation path from the starting point to the destination with the goal of minimizing the travel cost; Step B5, Output Path: Send the found optimal safe evacuation path to the user terminal for navigation guidance.
2. The urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights according to claim 1, characterized in that, Determine the data type of the multimodal data in order to use the corresponding processing method to obtain the water depth, specifically including: When the water depth reflected by the multimodal data is the water depth measured by the user through the tool, the water depth is read directly. When the water depth reflected by the multimodal data is a descriptive body part and its height, the body part-height ratio mapping table is looked up to obtain the ratio coefficient between the body part and the height, and the water depth is calculated based on the ratio coefficient and the height.
3. The urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights according to claim 1, characterized in that, Considering the timeliness and spatial consistency of the data, a dynamic overall credibility weight is calculated for each multimodal data point. Specifically, the overall credibility weight of the i-th multimodal data point is calculated according to the following formula. ; in, ; As a spatiotemporal decay weight used to characterize the timeliness of data, For multi-source verification weights used to characterize the consistency of the data space, The attenuation coefficient is... Let N be the difference between the reporting time of the i-th multimodal data point and the current time, and let N represent the number of neighboring points of the water accumulation point corresponding to the i-th multimodal data point. Z-score is used to characterize the difference between a given data point and the mean of its N neighboring data points. This is the preset Z-score threshold.
4. The urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights according to claim 1, characterized in that, Step A6 specifically includes: Road segment waterlogging data aggregation: For road segments Collect all reported locations located on or directly affected by this road section, forming a set of water accumulation information for that road section. ; Calculating the comprehensive assessment of water accumulation depth for road sections: A weighted average method is used, incorporating aggregate data. Water depth in all data and its overall credibility weight Calculate the road segment Comprehensive assessment of water depth : in, For the first The water depth of each data point; Determine the water depth weighting factor for road sections: Multiple water depth intervals are pre-set, and a water depth weighting factor is assigned to each interval, with a larger weighting factor for deeper water. This allows for the determination of the comprehensive assessment of water accumulation depth for the road section. The water depth weighting factor of the road section is determined by the water depth range it falls within. .
5. The urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights according to claim 1, characterized in that, The toll cost of each road segment in the global road network risk map is updated by combining the static attributes of the road and the water depth weight factor of the road segment. Specifically, the toll cost of road segment j is calculated according to the following formula. ; in, Representative road section Length, Representative road section Travel time per unit length under ideal conditions Indicates road segment The cost of passage under ideal conditions. Indicates road segment The water depth weighting factor is positively correlated with the water depth. When the water depth exceeds a threshold, Set to infinity make the road section The cost is infinitely high and therefore avoided; Indicates road segment The road width weighting factor is positively correlated with the road width; Indicates road segment The altitude weighting factor is positively correlated with altitude.
6. The urban flooding emergency route planning method based on multimodal crowdsourced data and dynamic risk weights according to claim 1, characterized in that, The path planning algorithm, with the goal of minimizing travel cost, searches for the optimal safe evacuation path from the starting point to the destination, specifically including: Starting from the user terminal's current location and ending at the designated evacuation target point, the calculated travel cost is used as the starting point. Using the weights of the edges corresponding to road segment j, the graph search algorithm is run to search for the path with the lowest travel cost in the global road network risk map, which is then used as the optimal safe evacuation path.
7. An urban flooding emergency route planning system based on multimodal crowdsourced data and dynamic risk weights, characterized in that, include: User terminals are used to report multimodal data reflecting water depth; Allows users to set evacuation target points and initiate requests; And receive the optimal safe evacuation route for navigation guidance; The application server is used to receive data packets from user terminals and perform format verification and cleaning on the data packets; determine the data type of multimodal data so as to use the corresponding processing method to obtain the water depth; unify the multimodal data reported by different user terminals into a standard format, and perform cross-validation on the relevant multimodal data within the spatial range; By considering the timeliness and spatial consistency of the comprehensive data, a dynamic comprehensive credibility weight is calculated for each piece of multimodal data. For each road segment in the road network, its related multimodal data is aggregated, and based on the comprehensive credibility weight, the comprehensive assessment water depth of the road segment is calculated, thereby determining the water depth weight factor of the road segment. Periodically calculate the comprehensive credibility weight of historical data based on the spatiotemporal decay model to achieve data aging processing and update the global road network risk map; When a user makes a request, the latest global road network risk map is obtained after being processed by the spatiotemporal decay model; Based on the latest global road network risk map, the passage cost of each road segment in the global road network risk map is updated by combining road static attributes and water depth weight factors of road segments; A spatial database is used to store multimodal data from user terminals and dynamic global road network risk maps; The route planning engine runs route planning algorithms to search for the optimal safe evacuation route from the starting point to the destination with the goal of minimizing travel costs, and then sends the found optimal safe evacuation route to the user terminal.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 6.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.