Real-time collection and fusion scheduling method for Internet of Things data of fusion communication platform

Abstract

The invention discloses a real-time collection and fusion scheduling method for Internet of Things data of a fusion communication platform, and relates to the technical field of data fusion and scheduling. The method comprises the steps of collecting multi-source traffic data, user travel demands and environmental factor data, and generating a travel route planning scheme through a multi-modal data fusion model. The model comprises a weather influence sub-model and an emergency response sub-model. Furthermore, according to the real-time position of the user and the travel emergency degree, the planning scheme weight is dynamically adjusted, and if a threshold value is exceeded, a transport capacity allocation or distribution instruction is triggered. Under the emergency conditions of rail transit delay and the like, an emergency scheduling mode is activated, delay duration, surrounding congestion and the number of affected passengers are comprehensively considered, a compensatory scheduling scheme is generated and comprises a connection bus scheduling strategy and a shared bicycle directional putting strategy, a connection route is optimized through a dynamic path planning method, and the evacuation efficiency is improved. According to the method, through multi-source data fusion and dynamic decision, the traffic scheduling accuracy and the emergency response capability are effectively improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of data fusion and scheduling, and in particular relates to a method for real-time collection and fusion scheduling of Internet of Things data on a fusion communication platform. Background Art

[0002] With the acceleration of urbanization and the rapid development of intelligent transportation systems, the real-time integration and dynamic scheduling of multi-source traffic data have become the core challenges in improving urban transportation efficiency. Current traffic management systems generally have the following technical bottlenecks:

[0003] Data silos and insufficient integration: Traditional systems lack effective integration mechanisms for processing heterogeneous data from multiple sources, including real-time public transportation arrival data, shared capacity distribution, and road network congestion information. Existing models struggle to dynamically adjust road network impedance parameters, especially in the event of sudden changes in weather conditions or emergencies. This can cause route planning to deviate from actual capacity. For example, most navigation systems rely solely on historical average speed estimates and fail to factor in real-time weather factors into attenuation coefficient calculations, resulting in inaccurate forecasts of transportation demand.

[0004] Dynamic Response Lag: Existing itinerary planning algorithms often use static weighting, failing to dynamically adjust priorities based on a user's real-time location or the urgency of their trip (e.g., medical emergencies, catching a flight). When a user deviates from the preset route or traffic conditions change suddenly, the system often requires manual re-planning, delaying optimal dispatch timing.

[0005] Rigid capacity allocation: Transport node capacity assessments often rely on fixed thresholds, lacking a dynamic perception mechanism for matching transport supply and demand. During peak hours, shuttle bus dispatch at subway stations often experience both localized overcapacity and shortages. This stems from a lack of real-time integration with dynamic capacity data, such as the availability of shared bikes and taxi density.

[0006] Lack of intelligent emergency dispatch: Existing emergency plans mostly adopt preset fixed strategies (such as adding a fixed number of shuttle buses) and lack the ability to make adaptive decisions based on the length of delays and the status of surrounding road networks.

[0007] Therefore, it is necessary to build a next-generation intelligent scheduling system with the following characteristics:

[0008] Deep fusion of multimodal data: Breaking through the limitations of a single data dimension, establishing a decision-making framework that integrates sub-models such as meteorological impact prediction and emergency response;

[0009] Dynamic evaluation system: Introducing a trip urgency quantification model and spatiotemporal attention mechanism to achieve millisecond-level dynamic weight allocation;

[0010] Flexible capacity matching: triggers graded warnings based on real-time supply and demand matching factors, supporting intelligent coordination of cross-regional capacity;

[0011] Self-healing emergency dispatch: Through time-varying road network modeling and connection priority weight calculation, a closed-loop compensatory dispatch chain is formed.

[0012] The invention is proposed to address the above-mentioned technical pain points. Through the innovative integration of real-time data collection of the Internet of Things, multimodal fusion modeling and dynamic path optimization algorithms, a full-chain technical solution from data perception to intelligent decision-making is constructed to enhance the resilient operation capabilities of urban transportation systems. Summary of the Invention

[0013] In order to overcome the shortcomings and deficiencies of the above-mentioned prior art, the present invention adopts the following technical solutions:

[0014] The real-time collection and fusion scheduling method of IoT data on the converged communication platform has the following process:

[0015] S10: Obtain multi-source traffic data, user travel demand data, and environmental factor data, and generate a travel route planning solution based on a multimodal data fusion model;

[0016] S20: Obtain the target user's current location and travel urgency parameters, generate a dynamic weight evaluation value based on the real-time location coordinates, travel urgency, and planning scheme, and if the weight evaluation value exceeds a preset threshold, proceed to S30; otherwise, maintain the original planning scheme;

[0017] S30: Obtain the connection capacity threshold and real-time passenger flow of each transportation node, and generate a capacity allocation instruction based on the node carrying capacity assessment model. The capacity allocation instruction includes an instruction to increase capacity or a diversion and guidance instruction;

[0018] S40: When a rail transit delay event is detected, the emergency dispatch mode is activated, and a compensatory dispatch plan is generated based on the delay duration and surrounding transportation resources. Preferably, the compensatory dispatch plan includes a shuttle bus dispatch strategy and a shared bicycle targeted deployment strategy.

[0019] Preferably, the multi-source traffic data includes real-time arrival data of public transportation, shared capacity distribution data and road network congestion data, the user travel demand data includes the departure place, destination and expected arrival time, and the environmental factor data includes meteorological data and event notification data; preferably, the multimodal data fusion model includes a weather impact sub-model and an emergency response sub-model, and the construction method of the weather impact sub-model includes collecting a matching data set of historical meteorological data and capacity adjustment records, training a random forest classifier to predict the capacity attenuation coefficient under different meteorological conditions, and the emergency response sub-model adopts an online learning mechanism to update the road network impedance matrix in real time when a traffic accident is detected; preferably, the construction method of the multimodal data fusion model includes establishing a traffic element feature vector, calculating the feature weight matrix through a spatiotemporal attention mechanism, and using a gated recurrent unit to perform multi-cycle feature fusion.

[0020] Preferably, the calculation formula of the travel urgency parameter E is:

[0021]

[0022] Among them, t remain is the remaining travel time, predicted based on real-time traffic conditions; t standard is the standard travel time, obtained based on the historical average; S is the safety factor, and 0≤S≤1;

[0023] When E<1, the emergency state is triggered and mapped to the [0,1] interval through the piecewise function:

[0024] Among them, k and λ are adjustment parameters.

[0025] Preferably, a comprehensive evaluation function including the time-space distance factor, the capacity supply-demand matching factor and the user preference coefficient is constructed. Specifically,

[0026] Preferably, the time-space distance factor is calculated by collecting real-time road network topology data, dynamic parameters of transportation tools, and coordinates of the starting and ending points of the user's trip, and obtaining the shortest time path in the real-time road network. Among them, l e is the length of the road section; v e (t) road speed limit; ρ e (t) is the congestion index (0≤ρ<1), and t Path Mapping to the interval [0,1]: Among them, d max Generate a spatiotemporal weight for the maximum possible travel distance of the road network combined with time sensitivity; Combined with time sensitivity τ, generate a spatiotemporal weight: spatiotemporal distance factor = τ t norm +(1-τ)·d norm ;t norm is the time normalized value, the method is the same as d norm ;

[0027] Preferably, the capacity supply and demand matching factor C is obtained by the calculation formula: Among them D t represents the forecast demand at time t; S t represents the available transport capacity at time t, which includes the density of public transportation schedules, the number of available shared vehicles, and the real-time distribution of taxis; when D t =S t When C=1, it means perfect match; when D t >>S t or D t ≤S tWhen C is less than 0.6, the capacity coordination warning is triggered, and when C is less than 0.4, cross-regional capacity scheduling is initiated.

[0028] Preferably, the user preference coefficient is updated by establishing a user travel mode feature library, including: a transportation tool selection tendency vector Q = (q1, q2, q3), where q1 represents the subway preference, q2 represents the bus preference, and q3 represents the shared travel preference; Among them, U similar is a similar user group; η is the learning rate; the time sensitivity parameter τ∈[0,1] and the cost sensitivity parameter λ∈[0,1] are used to update the user profile through the collaborative filtering algorithm. When a sudden change in the user behavior pattern is detected, a preference migration vector ΔQ is generated. High time sensitivity τ→1: path planning aims at the shortest time; high cost sensitivity λ→1: low-cost combinations are recommended first.

[0029] Calculate the comprehensive evaluation function: W = α·T + β·C + γ·P; where T represents the expected travel time deviation; C represents the matching degree of transport supply and demand; P represents the matching degree of user historical preferences; α, β, and γ are dynamic adjustment coefficients; dynamic thresholds are calculated using sliding windows. W t.24h It is the average of the comprehensive evaluation values ​​in the past 24 hours; is the standard deviation of the comprehensive evaluation value in the past 24 hours; in, is the rate of change of the comprehensive assessment value.

[0030] Preferably, the activation conditions of the emergency dispatch mode include: the rail transit delay duration exceeds the threshold T0; the surrounding road congestion index exceeds the threshold C0; the number of affected passengers reaches N; when two or more conditions are met at the same time, a level 3 emergency response is triggered.

[0031] Preferably, the compensatory scheduling scheme generating method includes:

[0032] Step 1: Calculate the radiation radius of the affected area R = min (R max ,kΔt); where R max is the maximum radiation radius, k is the radiation radius growth coefficient, and Δt is the rail transit delay duration;

[0033] Step 2: Determine the number of shuttle buses required Among them, P delay is the number of stranded passengers, C bus is the passenger capacity of each shuttle bus, Indicates rounding up;

[0034] Step 3: Optimize the placement of shared bicycles: Set up n temporary placement points within the radius r of the rail transit station to meet Among them, S i is the capacity of the ith temporary deployment point, n is the number of temporary deployment points, and μ is the single-bike connection ratio coefficient.

[0035] Preferably, the shuttle bus scheduling strategy includes a dynamic path planning method, and a time-varying road network model G(t) = (V, E, W(t)) is established; where V is the node set; E is the edge set; W(t) is the edge weight, and the edge weight W(t) is a comprehensive weight that changes with time and is determined by the real-time travel time weight ω t , connection priority weight ω p , line overlap penalty term ω o Composition; the edge weight calculation formula is W(t)=ω t +λ·ω p -μ·ω o ; Among them, λ is the connection priority coefficient. The larger the value, the higher the priority is for serving high-detention nodes. When λ>1, it emphasizes the rapid evacuation of key areas. μ is the line overlap penalty coefficient. The larger the value, the more forced decentralized scheduling.

[0036] Preferably, the real-time travel time weight ω t : Dynamic calculation based on current traffic status;

[0037] Preferably, the connection priority weight ω p : It is positively correlated with the number of stranded passengers. Priority is assigned based on the number of stranded passengers and the degree of urgency at the node. The more stranded passengers, the higher the priority.

[0038] Preferably, the line overlap penalty term ω o : It is positively correlated with the number of road sections covered by existing shuttle buses, avoiding multiple shuttle buses covering the same road section repeatedly and reducing resource waste.

[0039] Preferably, the scheduling rules of the dynamic path planning method are as follows:

[0040] During peak hours, increase λ and prioritize evacuation of stations with large passenger flows;

[0041] Road network congestion, improve μ, and avoid crowded sections of roads with shuttle buses;

[0042] Preferably, the dynamic path planning method includes:

[0043] Initialization: Set the initial cost of each node to infinity, the starting cost to 0, put the starting point in the priority queue, and mark it as unvisited;

[0044] Iterative search: extract the current minimum cost node, traverse adjacent nodes, calculate the dynamic edge weight from to , update the total cost of the node, and dynamically improve the search order of high-priority nodes;

[0045] Termination condition: Stop when the target node is marked as visited or the queue is empty.

[0046] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0047] 1. This invention integrates multi-source data, including real-time public transportation arrival data, shared capacity distribution data, road network congestion data, user travel needs, and meteorological and event notifications, to construct a weather impact sub-model and an emergency response sub-model. It then combines the spatiotemporal attention mechanism and gated recurrent units for feature fusion to form an accurate multimodal data fusion model. This model can predict the capacity attenuation coefficient under different meteorological conditions in real time and dynamically update the road network impedance matrix when a traffic accident occurs, thereby generating a more reasonable travel path planning solution. Compared with traditional static path planning, this improves the real-time and accuracy of traffic management and scheduling.

[0048] 2. The present invention introduces a comprehensive evaluation function including the time-space distance factor, the capacity supply and demand matching factor and the user preference coefficient, and dynamically adjusts the path planning strategy in combination with the travel urgency parameter. At the same time, an emergency dispatch mode is designed. When rail transit is delayed, a compensatory dispatch plan is automatically generated based on the delay duration, the surrounding congestion index and the number of affected passengers, including shuttle bus dispatch and shared bicycle targeted deployment. In emergency situations, such as rail transit delays, the system can quickly assess the affected area, intelligently dispatch shuttle buses and shared bicycles, effectively evacuate stranded passengers, and avoid passenger backlogs and traffic paralysis. The emergency response time is shortened;

[0049] 3. This invention constructs an adaptive adjustment strategy to dynamically adjust the spatiotemporal weights and user preference coefficient weights in the comprehensive evaluation function based on the urgency of the trip. Simultaneously, a collaborative filtering algorithm updates user profiles, captures sudden changes in user behavior patterns, and generates preference migration vectors to further optimize the alignment of route planning with user needs. This makes route planning more flexible, enabling it to prioritize time in high-urgency situations and quickly respond to urgent travel needs; while in low-urgency situations, it prioritizes user preferences and provides personalized travel solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 A flow chart showing the real-time collection and fusion scheduling method of the integrated communication platform IoT data of the present invention is shown;

[0052] Figure 2 A flow chart showing the generation of a compensatory scheduling scheme according to the present invention is shown;

[0053] Figure 3 A flow chart of dynamic path planning of the present invention is shown. DETAILED DESCRIPTION

[0054] 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 ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] In addition, the described features, structures or characteristics can be combined in any suitable manner in one or more example embodiments. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure can be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. can be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0056] Example 1:

[0057] See Figure 1 As shown, the real-time collection and fusion scheduling method of the converged communication platform IoT data of this embodiment has the following process:

[0058] S10: Acquire multi-source traffic data, user travel demand data, and environmental factor data, and generate a travel route planning solution based on a multimodal data fusion model. The multi-source traffic data includes real-time public transportation arrival data, shared transportation capacity distribution data, and road network congestion data; user travel demand data includes departure point, destination, expected arrival time, etc.; and environmental factor data includes meteorological data and event notification data.

[0059] The multimodal data fusion model includes a weather impact sub-model and an emergency response sub-model: the construction method of the weather impact sub-model includes: collecting a matching dataset of historical meteorological data and capacity adjustment records; training a random forest classifier to predict the capacity attenuation coefficient under different meteorological conditions; the emergency response sub-model adopts an online learning mechanism to update the road network impedance matrix in real time when a traffic accident is detected.

[0060] The multimodal data fusion model construction method includes:

[0061] a. Establish traffic element feature vector:

[0062]

[0063] Among them, v1 is the subway interval; v2 is the bus full load rate; v3 is the shared bicycle turnover rate; T max The longest subway interval in history; C max is the maximum passenger capacity of the bus; S max is the maximum turnover rate of shared bicycles.

[0064] b. Calculate the feature weight matrix through the spatiotemporal attention mechanism;

[0065] c. Use gated recurrent units for multi-cycle feature fusion.

[0066] S20: Obtain the current location and travel urgency parameters of the target user, generate a dynamic weight evaluation value based on the real-time location coordinates, travel urgency and planning scheme, and proceed to S30 if the weight evaluation value exceeds the preset threshold, otherwise maintain the original planning scheme.

[0067] Trip urgency parameter E,

[0068] Among them, t remain is the remaining travel time (predicted based on real-time traffic conditions); t standard is the standard travel time (historical average); S is the safety factor, and 0≤S≤1;

[0069] When E<1, the emergency state is triggered and mapped to the [0,1] interval through the piecewise function:

[0070] Among them, k and λ are adjustment parameters.

[0071] Adopting an adaptive adjustment strategy:

[0072] High urgency (E norm →1): α→1, focusing on the time dimension;

[0073] Low urgency (E norm→0): β→0.7, focusing on user preference.

[0074] Construct a comprehensive evaluation function that includes the time-space distance factor, the capacity supply and demand matching factor, and the user preference coefficient:

[0075] The space-time distance factor is obtained by:

[0076] Collect real-time road network topology data (road length, speed limit, congestion index), dynamic parameters of transportation tools (real-time arrival time of subway / bus, turnover speed of shared bicycles), and coordinates of the starting and ending points of user trips (latitude and longitude).

[0077] Calculate the shortest time path in real-time road network Among them, l e is the length of the road section; v e (t) road speed limit; ρ e (t) is the congestion index (0≤ρ<1).

[0078] t Path Mapping to the interval [0,1]: Among them, d max is the maximum possible travel distance of the road network.

[0079] Combined with the time sensitivity τ (user parameter), the spatiotemporal weight is generated: spatiotemporal distance factor = τ·t norm +(1-τ)·d norm ;t norm is the time normalized value, the method is the same as d norm .

[0080] The capacity supply and demand matching factor is obtained in the following way: The capacity supply and demand matching factor is calculated as follows: Among them D t represents the forecast demand at time t; S t represents the available transport capacity at time t, which includes the density of public transportation schedules, the number of available shared vehicles, and the real-time distribution of taxis; when D t =S t When C=1, it means perfect match; when D t >>S t or D t ≤S t When C is less than 0.6, the capacity coordination warning is triggered, and when C is less than 0.4, cross-regional capacity scheduling is initiated.

[0081] The user preference coefficient is obtained by: establishing a user travel mode feature library, which includes: a transportation tool selection tendency vector Q = (q1, q2, q3), where q1 represents the subway preference, q2 represents the bus preference, and q3 represents the shared travel preference;

[0082] renew: Among them, U similar is the similar user group; η is the learning rate.

[0083] The time sensitivity parameter τ∈[0,1] and the cost sensitivity parameter λ∈[0,1] are used to update the user profile through the collaborative filtering algorithm. When a sudden change in the user behavior pattern is detected (such as continuously choosing shared bicycles instead of subways), a preference migration vector ΔQ is generated.

[0084] High time sensitivity (τ→1): Path planning aims at the shortest time.

[0085] High cost sensitivity (λ→1): Prioritize low-cost combinations (such as bus + walking).

[0086] Calculate the comprehensive evaluation function: W = α·T + β·C + γ·P; where T is the deviation value of the expected travel time; C is the matching degree of transportation supply and demand; P is the matching degree of user historical preferences; α, β, and γ are dynamic adjustment coefficients.

[0087] Using sliding window statistics dynamic threshold W t.24h It is the average of the comprehensive evaluation values ​​in the past 24 hours; It is the standard deviation of the comprehensive evaluation value in the past 24 hours.

[0088] in, is the rate of change of the comprehensive assessment value.

[0089] S30: Obtain the connection capacity threshold and real-time passenger flow of each transportation node, and generate a capacity allocation instruction based on the node carrying capacity assessment model. The capacity allocation instruction includes an additional capacity instruction or a diversion guidance instruction.

[0090] S40: When a rail transit delay event is detected, an emergency dispatch mode is activated, and a compensatory dispatch plan is generated based on the delay duration and surrounding transport resources. The compensatory dispatch plan includes a shuttle bus dispatch strategy and a shared bicycle targeted deployment strategy.

[0091] The activation conditions of the emergency dispatch mode include:

[0092] The rail transit delay duration exceeds the threshold T0;

[0093] The surrounding road congestion index exceeds the threshold C0;

[0094] The number of affected passengers reaches N;

[0095] A level 3 emergency response is triggered when two or more conditions are met at the same time.

[0096] See Figure 2 As shown, the compensatory scheduling scheme generation method includes:

[0097] Step 1: Calculate the radiation radius of the affected area R = min (R max ,kΔt); where R max is the maximum radiation radius, k is the radiation radius growth coefficient, and Δt is the rail transit delay duration.

[0098] Step 2: Determine the number of shuttle buses required Among them, P delay is the number of stranded passengers, C bus is the passenger capacity of each shuttle bus, Indicates rounding up.

[0099] Step 3: Optimize the placement of shared bicycles: Set up n temporary placement points within the radius r of the rail transit station to meet Among them, S i is the capacity of the ith temporary deployment point, n is the number of temporary deployment points, and μ is the single-bike connection ratio coefficient.

[0100] The shuttle bus scheduling strategy includes a dynamic path planning method: establishing a time-varying road network model G(t) = (V, E, W(t)); where V is the node set; E is the edge set; W(t) is the edge weight, and the edge weight W(t) is a comprehensive weight that changes over time and consists of the following three parts: real-time travel time weight ω t , connection priority weight ω p , line overlap penalty term ω o ;

[0101] Real-time travel time weight ω t : Dynamic calculation based on current traffic status (congestion index, accidents).

[0102] Connection priority weight ω p : It is positively correlated with the number of stranded passengers and is assigned priority based on the number of stranded passengers and the degree of urgency at the node (the more stranded passengers, the higher the priority).

[0103] Line overlap penalty term ω o : It is positively correlated with the number of road sections covered by existing shuttle buses, avoiding multiple shuttle buses covering the same road section repeatedly and reducing resource waste.

[0104] Edge weight calculation formula: W(t) = ω t +λ·ω p -μ·ω o ; Among them, λ is the connection priority coefficient. The larger the value, the higher the priority is for serving high-detention nodes (when λ>1, it emphasizes the rapid evacuation of key areas); μ is the line overlap penalty coefficient. The larger the value, the more forced decentralized scheduling.

[0105] The scheduling rules are as follows:

[0106] During peak hours, increase λ and prioritize evacuation of stations with large passenger flows;

[0107] The road network is congested, so we should improve μ to avoid shuttle buses crowding the roads.

[0108] See Figure 3 As shown in Figure 2, the process of the dynamic path planning method is as follows:

[0109] a. Initialization: Set the initial cost of each node to infinity and the starting cost to 0. Put the starting point in the priority queue and mark it as unvisited.

[0110] b. Iterative search:

[0111] Extract the current minimum cost node μ: Take the node with the minimum cost from the queue.

[0112] Traverse the adjacent node ν and calculate the dynamic edge weight W from μ to ν μν (t).

[0113] The total cost of updating node ν: Cost = min(Cost(ν), Cost(μ) + W μν (t));

[0114] Priority adjustment: According to ω p Dynamically promote the search order of high priority nodes.

[0115] c. Termination condition: Stop when the target node (connection endpoint) is marked as visited or the queue is empty.

[0116] For example:

[0117] Scenario description: The subway is delayed for 30 minutes, 200 passengers are stranded, and the surrounding road congestion index is ρ e (t) = 0.8 (the threshold is 0.7). Activate the Level 3 emergency response, and dispatch shuttle buses to evacuate stranded passengers.

[0118] Dynamic path planning process:

[0119] a. Constructing a time-varying road network:

[0120] Nodes: subway station A, temporary transfer point B / C, bus hub D.

[0121] Edge weights:

[0122] ω t : Flooding on the road causes the travel time from A to B to increase by 50%.

[0123] ω p:200 people stranded at subway station A →ω p =3.

[0124] ω o :There are 2 buses covering A→B→ω o =2.

[0125] b. Algorithm calculation:

[0126] Optimal path: Subway station A → Bus hub D (minimum cost: W = 1.2 + 1.5 × 3 - 0.8 × 0 = 5.7).

[0127] Suboptimal path: subway station A → temporary transfer point B (cost: W = 2.0 + 1.5 × 3 - 0.8 × 2 = 5.9W = 2.0 + 1.5 × 3 - 0.8 × 2 = 5.9).

[0128] c. Scheduling decision: Path A→D is preferred, while spare vehicles are reserved to deal with sudden congestion.

[0129] The beneficial effects of this embodiment include: improving path planning efficiency and emergency response capabilities through multi-source data fusion and dynamic scheduling, achieving optimal allocation of traffic resources, enhancing user experience, and reducing operational risks.

[0130] The weights of this invention are used to measure the degree to which different factors or variables influence a particular outcome or decision. Weights are defined as numerical values ​​assigned to each factor when comparing and evaluating multiple factors, reflecting its importance or priority. These weights can be determined based on specific circumstances and needs, and are typically developed and confirmed by professionals or stakeholders. By properly setting weights, programs or systems can be made to make more accurate decisions or predictions.

[0131] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

[0132] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for real-time collection and fusion scheduling of IoT data on a converged communication platform, characterized in that: The method flow is as follows: S10: Obtain multi-source traffic data, user travel demand data, and environmental factor data, and generate a travel route planning solution based on a multimodal data fusion model; S20: Obtain the target user's current location and travel urgency parameters, generate a dynamic weight evaluation value based on the real-time location coordinates, travel urgency, and planning scheme, and if the weight evaluation value exceeds a preset threshold, proceed to S30; otherwise, maintain the original planning scheme; S30: Obtain the connection capacity threshold and real-time passenger flow of each transportation node, and generate a capacity allocation instruction based on the node carrying capacity assessment model. The capacity allocation instruction includes an instruction to increase capacity or a diversion and guidance instruction; S40: When a rail transit delay event is detected, the emergency dispatch mode is activated, and a compensatory dispatch plan is generated based on the delay duration and surrounding transportation resources. The compensatory dispatch plan includes a shuttle bus dispatch strategy and a shared bicycle targeted deployment strategy.

2. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 1, characterized in that: The multi-source traffic data includes real-time public transportation arrival data, shared capacity distribution data and road network congestion data; user travel demand data includes departure place, destination and expected arrival time; environmental factor data includes meteorological data and event notification data; the multimodal data fusion model includes a weather impact sub-model and an emergency response sub-model; the construction method of the weather impact sub-model includes collecting a matching data set of historical meteorological data and capacity adjustment records, training a random forest classifier to predict the capacity attenuation coefficient under different meteorological conditions; the emergency response sub-model adopts an online learning mechanism to update the road network impedance matrix in real time when a traffic accident is detected; the construction method of the multimodal data fusion model includes establishing a traffic element feature vector, calculating a feature weight matrix through a spatiotemporal attention mechanism, and using a gated recurrent unit to perform multi-cycle feature fusion.

3. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 1, characterized in that: The calculation formula of the trip urgency parameter E is: Among them, t remain is the remaining travel time, predicted based on real-time traffic conditions; t standard is the standard travel time, obtained based on the historical average; S is the safety factor, and 0≤S≤1; When E<1, the emergency state is triggered and mapped to the [0,1] interval through the piecewise function: Among them, k and λ are adjustment parameters.

4. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 1, characterized in that: Construct a comprehensive evaluation function that includes the time-space distance factor, the capacity supply and demand matching factor, and the user preference coefficient. Specifically, The time-space distance factor is calculated by collecting real-time road network topology data, dynamic parameters of transportation tools, and coordinates of the starting and ending points of the user's trip, and obtaining the shortest time path in the real-time road network. Among them, l e is the length of the road section; v e (t) road speed limit; ρ e (t) is the congestion index (0≤ρ<1), and t Path Mapping to the interval [0,1]: Among them, d max Generate a spatiotemporal weight for the maximum possible travel distance of the road network combined with time sensitivity; Combined with time sensitivity τ, generate a spatiotemporal weight: spatiotemporal distance factor = τ t norm +(1-τ)·d norm ;t norm is the time normalized value, the method is the same as d norm ; The capacity supply and demand matching factor C is obtained by the calculation formula: Among them D t represents the forecast demand at time t; S t represents the available transport capacity at time t, which includes the density of public transportation schedules, the number of available shared vehicles, and the real-time distribution of taxis; when D t =S t When C=1, it means perfect match; when D t >>S t or D t ≤S t When C is less than 0.6, the capacity coordination warning is triggered, and when C is less than 0.4, cross-regional capacity scheduling is initiated. The user preference coefficient is updated by establishing a user travel mode feature library, including: a transportation tool selection tendency vector Q = (q1, q2, q3), where q1 represents the subway preference, q2 represents the bus preference, and q3 represents the shared travel preference; Among them, U similar is a similar user group; η is the learning rate; the time sensitivity parameter τ∈[0,1] and the cost sensitivity parameter λ∈[0,1] are used to update the user profile through the collaborative filtering algorithm. When a sudden change in the user behavior pattern is detected, a preference migration vector ΔQ is generated. High time sensitivity τ→1: path planning aims at the shortest time; high cost sensitivity λ→1: low-cost combinations are recommended first. Calculate the comprehensive evaluation function: W = α·T + β·C + γ·P; where T represents the expected travel time deviation; C represents the matching degree of transport supply and demand; P represents the matching degree of user historical preferences; α, β, and γ are dynamic adjustment coefficients; dynamic thresholds are calculated using sliding windows. W t.24h It is the average of the comprehensive evaluation values ​​in the past 24 hours; is the standard deviation of the comprehensive evaluation value in the past 24 hours; in, is the rate of change of the comprehensive assessment value.

5. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 1, characterized in that: The activation conditions of the emergency dispatch mode include: the rail transit delay duration exceeds the threshold T0; the surrounding road congestion index exceeds the threshold C0; the number of affected passengers reaches N; when two or more conditions are met at the same time, a level 3 emergency response is triggered.

6. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 1, characterized in that: The compensatory scheduling scheme generation method includes: Step 1: Calculate the radiation radius of the affected area R = min (R max ,kΔt); where R max is the maximum radiation radius, k is the radiation radius growth coefficient, and Δt is the rail transit delay duration; Step 2: Determine the number of shuttle buses required Among them, P delay is the number of stranded passengers, C bus is the passenger capacity of each shuttle bus, Indicates rounding up; Step 3: Optimize the placement of shared bicycles: Set up n temporary placement points within the radius r of the rail transit station to meet Among them, S i is the capacity of the ith temporary deployment point, n is the number of temporary deployment points, and μ is the single-bike connection ratio coefficient.

7. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 1, characterized in that: The shuttle bus scheduling strategy includes a dynamic path planning method, which establishes a time-varying road network model G(t) = (V, E, W(t)); where V is the node set; E is the edge set; W(t) is the edge weight, and the edge weight W(t) is a comprehensive weight that changes with time and is determined by the real-time travel time weight ω t , connection priority weight ω p , line overlap penalty term ω o Composition; the edge weight calculation formula is W(t)=ω t +λ·ω p -μ·ω o ; Among them, λ is the connection priority coefficient. The larger the value, the higher the priority is for serving high-detention nodes. When λ>1, it emphasizes the rapid evacuation of key areas. μ is the line overlap penalty coefficient. The larger the value, the more forced decentralized scheduling.

8. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 7, characterized in that: The real-time travel time weight ω t : Dynamic calculation based on current traffic status; The connection priority weight ω p : It is positively correlated with the number of stranded passengers. Priority is assigned based on the number of stranded passengers and the degree of urgency at the node. The more stranded passengers, the higher the priority. The line overlap penalty term ω o : It is positively correlated with the number of road sections covered by existing shuttle buses, avoiding multiple shuttle buses covering the same road section repeatedly and reducing resource waste.

9. The method for real-time collection and fusion scheduling of IoT data on a fusion communication platform according to claim 7, characterized in that: The scheduling rules of the dynamic path planning method are as follows: During peak hours, increase λ and prioritize evacuation of stations with large passenger flows; Road network congestion, improve μ, and avoid crowded sections of roads with shuttle buses; The dynamic path planning method comprises: Initialization: Set the initial cost of each node to infinity, the starting cost to 0, put the starting point in the priority queue, and mark it as unvisited; Iterative search: extract the current minimum cost node, traverse adjacent nodes, calculate the dynamic edge weight from to , update the total cost of the node, and dynamically improve the search order of high-priority nodes; Termination condition: Stop when the target node is marked as visited or the queue is empty.

Images