Intelligent collection and transportation scheduling method and system for urban construction waste based on internet of things

Abstract

The application discloses a city construction waste intelligent cleaning and transporting scheduling method and system based on Internet of Things, and contents include construction of basic construction waste data set, waste generation demand prediction, task allocation and cleaning and transporting scheduling. The application relates to the technical field of city construction waste intelligent management, and particularly relates to a city construction waste intelligent cleaning and transporting scheduling method and system based on Internet of Things. Through trend decomposition, multi-scale time sequence coding and full-load time prediction, the scheme combines road traffic states to construct cleaning and transporting urgency, and improves the accuracy and foresight of demand identification. Then, task allocation is performed according to demand intensity, vehicle position, load and regional pressure, and the optimization result is repaired through constraint to reduce repeated vehicle dispatching. Finally, a cleaning and transporting path is generated according to scheduling priority, and re-scheduling is triggered in combination with real-time states to enhance the real-time performance, adaptability and executability of the cleaning and transporting process.

Description

Technical Field

[0001] This invention relates to the field of intelligent management technology for urban construction waste, specifically to an intelligent urban construction waste collection and dispatching method and system based on the Internet of Things. Background Technology

[0002] With the accelerating pace of urbanization, construction, renovation, demolition, and municipal renewal activities are increasing, resulting in a large-scale, widely distributed, and highly fluctuating volume of urban construction waste. Simultaneously, the development of IoT technology, vehicle-mounted terminals, road sensing equipment, and data analysis algorithms provides a more refined, real-time, and intelligent technological foundation for construction waste management. Currently, construction waste management is gradually shifting from traditional manual experience-based scheduling and fixed-route transportation to data-driven, dynamic prediction, and intelligent optimization. In the future, it is expected to achieve collaborative management of the entire process of waste generation, vehicle status, road traffic, and task execution, thereby improving urban waste management efficiency, resource utilization, and management precision.

[0003] However, existing construction waste removal technologies still have significant shortcomings. On the one hand, many solutions rely mainly on static statistics or coarse-grained empirical rules to determine the amount of waste generated and the time of full load, which makes it difficult to reflect changes in construction activities, time-period fluctuations, and sudden accumulation situations in a timely manner, resulting in delayed prediction results and untimely removal responses. On the other hand, existing scheduling methods usually only consider distance, number of vehicles, or simple load constraints, lacking a comprehensive analysis of the urgency of locations, road traffic conditions, task relationships, and vehicle execution conflicts. This can easily lead to problems such as uneven task allocation, unreasonable route planning, high vehicle empty running rates, and difficulty in adjusting scheduling results in real time. Overall coordination and executability still need to be improved. Summary of the Invention

[0004] To address the above issues and overcome the shortcomings of existing technologies, this invention provides an intelligent urban construction waste collection and scheduling method and system based on the Internet of Things (IoT). Addressing the problems in existing urban construction waste collection schemes, such as insufficient understanding of waste generation trends, difficulty in timely identification of sudden increases and full-load risks, leading to delayed demand forecasting and untimely collection responses, this solution performs trend decomposition, multi-scale time-series coding, and full-load time prediction on multi-source collected data. Furthermore, it integrates road traffic conditions to construct a collection urgency index, enabling the model to more accurately identify the actual collection demand and time sensitivity at each location. Moreover, it addresses the shortcomings of existing technologies that often only consider single distance or static capacity factors in collection task allocation, lacking collaborative analysis of demand intensity, spatial correlation, and vehicle carrying capacity. To address the issues of fixed route planning, inability to dynamically adjust routes based on road conditions and task changes, and the tendency for overloading, timeouts, and execution conflicts in existing construction waste removal processes, this solution constructs scheduling priorities, generates removal routes that meet load and continuous service requirements, and monitors road conditions, vehicle remaining load, and new waste additions in real time during execution. This triggers a rolling rescheduling mechanism to rearrange unexecuted tasks, making the removal process more real-time and adaptable.

[0005] The technical solution adopted in this invention is as follows: an intelligent urban construction waste collection and scheduling method based on the Internet of Things, which includes the following steps:

[0006] Step S1: Construct a basic construction waste dataset by synchronously collecting multi-source information from waste generation points, road sensing devices, and vehicle terminals, and uniformly aligning data of different frequencies and formats. At the same time, complete outlier correction and standardized coding to form a dataset that can be directly used for subsequent prediction and scheduling.

[0007] Step S2: Waste generation demand forecasting. Forecast the future collection demand of each construction waste site; first extract the stable trend part, then separate the real demand fluctuation, and then combine multi-scale time series characteristics and time period information to predict the future waste growth, full load time and collection urgency, so as to obtain the demand intensity and priority service basis for each site.

[0008] Step S3: Task allocation. Based on the prediction results, the task allocation for waste removal is completed. First, the demand intensity, spatial location and road conditions of each point are combined to build a correlation. Then, high demand points are selected as regional centers and expanded to form waste removal areas. Subsequently, the vehicles are matched with the areas based on vehicle location, load capacity and regional demand. At the same time, unreasonable allocations such as conflicts and overloading are corrected to obtain an executable task allocation.

[0009] Step S4: Waste collection and dispatching. The specific waste collection and dispatching of the assigned tasks are carried out. The dispatching priority is generated based on the urgency of the location, the predicted amount of waste and the time of full load. Then, a waste collection route that meets the load and return requirements is planned for each vehicle.

[0010] Further, in step S1, constructing the basic construction waste dataset specifically includes the following steps:

[0011] Step S11: Data acquisition. Data is collected synchronously from construction waste generation points, road sensing devices, and vehicle terminals. Data of different frequencies and formats are unified to the same time granularity to form the original observation vector that can be directly called for subsequent modeling.

[0012] Step S12: Outlier identification, correction of outliers, drift values ​​and missing values ​​that appear during the collection process; for outliers, first construct correction weights based on neighboring sites and adjacent time points, and then replace the original outliers with spatiotemporal consistency estimates.

[0013] Step S13: Standardize the coding, transform the corrected multi-source data into time window samples of a uniform scale, and explicitly add time phase information so that the model can identify daily, weekly, and construction rhythm changes.

[0014] Furthermore, in step S2, the waste generation demand forecasting specifically includes the following steps:

[0015] Step S21: Construct a trend baseline. First, extract a slowly changing trend baseline from historical waste disposal data. Then, use the difference between the current observation and the trend baseline to represent the actual demand fluctuations, separating stable growth from sudden growth.

[0016] Step S22: Multi-scale temporal coding. Multi-scale convolutional coding is performed on the residual sequence, and the response at different scales is dynamically gated using temporal phase information, so that the model can capture both short-term surges caused by construction activities and slow rises caused by cycles; then attention-weighted fusion is performed on the outputs at different scales.

[0017] Step S23: Predict the full load time, predict the future amount of garbage growth, and add the growth amount to the current occupied capacity to directly deduce the full load time; first, predict the future growth amount; then, recursively obtain the future capacity occupancy; finally, calculate the full load time.

[0018] Step S24: Introduce constraints, integrate the growth rate, full load time, and road traffic status into a waste removal urgency index, construct a joint training loss, and add physical consistency constraints during the training phase.

[0019] Furthermore, in step S3, the task allocation specifically includes the following steps:

[0020] Step S31: Calculate the site association strength. Convert the predicted results of each construction waste site into a collection demand strength value, and construct the site association strength by combining spatial distance, road resistance and urgency. First, calculate the demand strength value of the site; then construct the association strength between any two sites.

[0021] Step S32: Initial area division. Select high-demand collection points from all construction waste collection points and form an initial collection area with these high-demand collection points as the core. First, select the point with the greatest collection demand as the center point of the area.

[0022] Step S33: Area expansion, incorporating non-central points into the most suitable area, so that the points within each area are consistent in terms of the intensity of waste collection demand and road accessibility;

[0023] Step S34: Task matching. First, calculate the task matching score of the vehicle for each region based on the vehicle's current location, remaining load, travel cost, and regional pressure value. The higher the matching score, the more suitable the vehicle is for undertaking the task in that region. Then, perform initial matching of the vehicle and the region according to the principle of maximizing the matching score. After the initial matching is completed, check the capacity constraints, task uniqueness constraints, and road conflict constraints. If it is found that a vehicle is overloaded with tasks, multiple vehicles are repeatedly assigned to the same region, or there are obvious conflicts between task paths, the allocation results are readjusted through the repair mechanism until all tasks meet the executable conditions.

[0024] Furthermore, in step S4, the waste collection and scheduling specifically includes the following steps:

[0025] Step S41: Schedule priority construction, which will be carried out in step S41. The output vehicle and area matching results are transformed into a sequence of cleanup tasks to be executed. The scheduling priority is constructed by combining the urgency of the location, the predicted amount to be cleaned up, and the estimated full load time, so that high-urgency and high-risk locations can obtain cleanup resources first.

[0026] Step S42: Generation of collection routes. For each vehicle, under the conditions of satisfying the maximum waste carrying capacity constraint, service continuity constraint, and return to the depot constraint, the optimal collection route is generated, and joint optimization is performed with the objectives of minimizing the total travel cost, minimizing the timeout risk, and maximizing the priority benefit; constraints are introduced in the optimization process.

[0027] Step S43: Rescheduling. During the vehicle's collection process, the road traffic status, vehicle remaining load, and new garbage growth information are monitored in real time. When a significant change is detected in the task execution conditions, a rolling re-optimization mechanism is triggered to reschedule tasks that have not yet been executed.

[0028] The present invention provides an intelligent urban construction waste collection and dispatching system based on the Internet of Things, which includes a basic construction waste dataset construction module, a waste generation demand prediction module, a task allocation module, and a collection and dispatching module.

[0029] The module for constructing a basic construction waste dataset synchronously collects multi-source information from waste generation points, road sensing devices, and vehicle terminals, and uniformly aligns data of different frequencies and formats. It also completes outlier correction and standardized coding to form a dataset that can be directly used for subsequent prediction and scheduling, and sends the data to the waste generation demand prediction module.

[0030] The waste generation demand prediction module receives data sent by the basic construction waste dataset construction module and predicts the future collection demand of each construction waste site. It first extracts the stable trend part, then separates the real demand fluctuations, and then combines multi-scale time series features and time period information to predict the future waste growth, full load time and collection urgency, obtains the demand intensity and priority service basis for each site, and sends the data to the task allocation module.

[0031] The task allocation module receives data sent by the waste generation demand prediction module and completes the allocation of collection tasks based on the prediction results. First, it integrates the demand intensity, spatial location and road conditions of each location to build a correlation. Then, it selects high-demand points as regional centers and expands them to form collection areas. Subsequently, it combines vehicle location, load capacity and regional demand to match vehicles with areas, while correcting unreasonable allocations such as conflicts and overloading, to obtain an executable task allocation and send the data to the collection scheduling module.

[0032] The waste collection and dispatch module receives data sent by the task allocation module and performs specific waste collection and dispatch for the already allocated tasks. It generates dispatch priorities based on the urgency of the location, the predicted amount of waste, and the time to full load, and then plans a waste collection route for each vehicle that meets the load and return requirements.

[0033] The beneficial effects achieved by the present invention using the above solution are as follows:

[0034] (1) In view of the problem that the existing urban construction waste removal schemes do not have a good grasp of the changing trend of waste generation and are difficult to identify sudden growth and full load risks in a timely manner, resulting in delayed demand forecasting and untimely removal response, this scheme decomposes multi-source data into trends, performs multi-scale time-series coding and full load time prediction, and further integrates road traffic status to construct a removal urgency index, so that the model can more accurately identify the real removal demand and time sensitivity of each location, thereby improving the accuracy and foresight of waste generation demand forecasting and providing a reliable basis for subsequent scheduling.

[0035] (2) In view of the problem that the allocation of waste collection tasks in the existing technology often only considers a single distance or static capacity factor and lacks collaborative analysis of demand intensity, spatial correlation and vehicle carrying capacity, this solution constructs the point demand intensity and station correlation intensity, first divides the high demand points into regions, and then combines vehicle location, remaining load, driving cost and regional pressure to match tasks. The allocation results are optimized through capacity constraints, uniqueness constraints and conflict repair mechanisms, so that the task allocation is more balanced and reasonable, which can improve vehicle utilization and reduce duplicate dispatch and resource waste.

[0036] (3) In response to the problems of fixed route planning, inability to dynamically adjust according to road conditions and task changes, and easy occurrence of overloading, timeout and execution conflict in the existing construction waste transportation process, this solution constructs scheduling priorities, generates transportation routes that meet the requirements of load capacity and continuous service, and monitors the road traffic status, vehicle remaining load and new waste changes in real time during the execution process, triggering a rolling rescheduling mechanism to rearrange unexecuted tasks, making the transportation process more real-time and adaptable, reducing the risk of overloading and delay, and improving the overall transportation efficiency and scheduling feasibility. Attached Figure Description

[0037] Figure 1 A schematic diagram of the intelligent urban construction waste collection and scheduling method based on the Internet of Things provided by the present invention;

[0038] Figure 2 A schematic diagram of the intelligent urban construction waste collection and dispatching system based on the Internet of Things provided by the present invention;

[0039] Figure 3 This is a schematic diagram of step S1;

[0040] Figure 4 This is a schematic diagram of step S2;

[0041] Figure 5 This is a schematic diagram of step S3;

[0042] Figure 6 This is a schematic diagram of step S4.

[0043] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0044] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0045] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0046] Example 1, see Figure 1 The present invention provides an intelligent urban construction waste collection and scheduling method based on the Internet of Things, which includes the following steps:

[0047] Step S1: Construct a basic construction waste dataset by synchronously collecting multi-source information from waste generation points, road sensing devices, and vehicle terminals, and uniformly aligning data of different frequencies and formats. At the same time, complete outlier correction and standardized coding to form a dataset that can be directly used for subsequent prediction and scheduling.

[0048] Step S2: Waste generation demand forecasting. Forecast the future collection demand of each construction waste site; first extract the stable trend part, then separate the real demand fluctuation, and then combine multi-scale time series characteristics and time period information to predict the future waste growth, full load time and collection urgency, so as to obtain the demand intensity and priority service basis for each site.

[0049] Step S3: Task allocation. Based on the prediction results, the task allocation for waste removal is completed. First, the demand intensity, spatial location and road conditions of each point are combined to build a correlation. Then, high demand points are selected as regional centers and expanded to form waste removal areas. Subsequently, the vehicles are matched with the areas based on vehicle location, load capacity and regional demand. At the same time, unreasonable allocations such as conflicts and overloading are corrected to obtain an executable task allocation.

[0050] Step S4: Waste collection and dispatching. The specific waste collection and dispatching of the assigned tasks are carried out. The dispatching priority is generated based on the urgency of the location, the predicted amount of waste and the time of full load. Then, a waste collection route that meets the load and return requirements is planned for each vehicle.

[0051] Example 2, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S1, the construction of the basic construction waste dataset specifically includes the following steps:

[0052] Step S11: Data Acquisition. Data is collected synchronously from construction waste generation points, road sensing devices, and vehicle terminals. Data of different frequencies and formats are unified to the same time granularity to form raw observation vectors that can be directly used for subsequent modeling. Specifically, at each sampling time, the first... The amount of waste, full load status, location coordinates, road traffic status, vehicle location, and vehicle load status of each construction waste site are used to construct an observation vector: ; and according to a uniform sampling interval Perform time resampling to obtain the aligned time index: ;in, Indicates the construction waste collection point number. This represents the unified discrete-time index. Represents the original continuous time. Indicates the first Each point at time The original observation vector, Indicates the start time of data collection. Indicates a uniform sampling interval. Indicates the first Each point at time The amount of garbage, This represents the full load value, taken as the full load rate. Represents the position coordinate vector. Represents the road traffic state vector. Represents the vehicle position vector. Indicates the vehicle's load status. This indicates the transpose operation. Indicates rounding down;

[0053] Step S12: Outlier identification, correcting outliers, drift values, and missing values ​​that occurred during the data collection process; for outliers, first construct correction weights based on neighboring stations and adjacent time points, then replace the original outliers with spatiotemporal consistency estimates, as shown below:

[0054] ;

[0055] in, Indicates the first The point and the first Each neighboring point at time Correction weights, This represents the spatial distance between two points. Indicates the distance attenuation scale. This indicates the rate of decay of differences in waste volume. This represents the corrected data vector. This represents a validity mask vector, where an element with a value of 1 indicates that the data in that dimension is valid, and a value of 0 indicates that the data in that dimension needs to be corrected. This represents element-wise multiplication. Indicates the first The spatial neighborhood set of each point Indicates the time smoothing weight, Represents the data vector from the previous moment;

[0056] Step S13: Standardization coding transforms the corrected multi-source data into time window samples of a uniform scale, and explicitly incorporates time phase information, enabling the model to identify daily, weekly, and construction rhythm variations; a standardized feature vector is constructed for each point. ; and will recently The feature combination at each time step is used to predict the input sequence: ;in, This represents the set of time-window samples used for subsequent demand forecasting. Indicates the first Each point at time The standardized feature vector, and These represent the mean and standard deviation of the amount of waste, respectively. and These represent the mean and standard deviation under full load conditions, respectively. and These represent the mean and standard deviation of the vehicle under load conditions, respectively. This represents the time phase vector, used to characterize hourly, weekday, and holiday cycles. This indicates the value for holidays; a value of 1 is used for holidays, and a value of 0 is used for non-holidays. and These represent hourly and weekday indexes, respectively. Indicates the length of the time window. This represents the set of time window samples used for subsequent demand forecasting.

[0057] Example 3, see Figure 1 and Figure 4 This embodiment is based on the above embodiment. In step S2, the waste generation demand prediction specifically includes the following steps:

[0058] Step S21: Construct a trend baseline. First, extract a slowly changing trend baseline from historical waste disposal data. Then, use the difference between the current observation and the trend baseline to represent the actual demand fluctuations, separating stable growth from sudden increases. Construct a recursive baseline from each point: Then calculate the residual vector: ;in, Indicates the first Each point at time The trend baseline vector, Represents the baseline smoothing coefficient and satisfies , Represents the residual vector;

[0059] Step S22: Multi-scale temporal coding. Multi-scale convolutional coding is performed on the residual sequence, and the responses at different scales are dynamically gated using temporal phase information. This enables the model to capture both short-term spikes caused by construction activities and slow rises caused by cycles. The temporal coding of each scale is defined as follows: Then, attention-weighted fusion is performed on the outputs at different scales: ;in, Indicates the first Encoding results at each scale express function, and They represent the first The linear transformation matrix and bias terms corresponding to each scale This represents vector concatenation. Indicates the expansion factor as Temporal convolutional network operators, Indicates the total number of scales. Indicates the first Attention weights at each scale Represents the attention query vector. Represents the attention transformation matrix. This represents the fused temporal feature vector;

[0060] Step S23: Predict the full load time, predict the future amount of garbage growth, and add the growth amount to the current occupied capacity to directly extrapolate the full load time; first, predict the future... Predict the growth rate of steps: Then, the future capacity occupancy is calculated recursively: Finally, the full load time was calculated: ;in, Indicates the first The future location of the first point The step-by-step prediction of waste growth, express function, Represents the growth rate mapping vector. Represents the step-size phase mapping vector. This represents the bias term mapping the growth amount. Indicates the future number Step size encoding, This indicates the currently used capacity. Indicates prediction up to the 1st Cumulative capacity occupied during the step Indicates the maximum prediction step size. Indicates the total capacity of the points. This indicates the predicted full load time, if within the prediction step size. If the interior is not fully loaded, then... As a flag indicating that the load is not full;

[0061] Step S24: Introduce constraints, integrating growth rate, full load time, and road traffic conditions into a waste collection urgency index, and add physical consistency constraints during the training phase. The waste collection urgency is defined as: The joint training loss is defined as: ; ;in, Indicates the first The urgency index of waste collection at each location, , , Represents the fusion coefficient. This indicates the urgency bias term. Represents the overall loss function. This indicates the total number of construction waste collection points. and These represent the projected growth and the actual growth, respectively. Indicates the actual full load time. Indicates the true urgency level. This represents the binary cross-entropy loss function. , , and Indicates the loss weight. Indicates the loss of physical consistency. This indicates taking the maximum value.

[0062] By performing the above operations, this solution addresses the problems in existing urban construction waste removal schemes, such as insufficient understanding of waste generation trends, difficulty in timely identification of sudden increases and full-load risks, leading to lagging demand forecasting and untimely removal responses. This solution performs trend decomposition, multi-scale time-series coding, and full-load time prediction on multi-source collected data, and further integrates road traffic conditions to construct a removal urgency index. This enables the model to more accurately identify the actual removal demand and time sensitivity at each location, thereby improving the accuracy and foresight of waste generation demand forecasting and providing a reliable basis for subsequent scheduling.

[0063] Example 4, see Figure 1 and Figure 5 This embodiment is based on the above embodiment. In step S3, the task allocation specifically includes the following steps:

[0064] Step S31: Calculate the site association strength. Transform the predicted results for each construction waste site into a collection demand intensity value, and combine this with spatial distance, road resistance, and urgency to construct the site association strength. First, calculate the... Demand intensity values ​​for each location: Then construct the correlation strength between any two points: ;in, Indicates the first The intensity of waste removal demand at each location. , and Indicates the strength fusion coefficient. This represents the average predicted growth rate for all data points at the current moment. Indicates point with point The strength of the correlation between them This represents the Euclidean distance between point i and point j. Indicates the distance attenuation scale. This represents the road impedance value between two points. Indicates the road impedance attenuation scale. , and Indicates the weight of related items. This indicates the scale of the decline in the intensity of demand.

[0065] Step S32: Initial area division. Select high-demand collection points from all construction waste collection points, and form an initial collection area with these high-demand collection points as the core. First, select the point with the highest collection demand as the center point of the area, as shown below:

[0066] ;

[0067] in, Indicates the first The central point of each region, Indicates the preceding A set of center points of the selected region Indicates the separation penalty coefficient. Indicates the central separation scale. Indicates a region index. Indicates the number of regions. This indicates taking the maximum value from the set of candidate points;

[0068] Step S33: Area expansion, incorporating non-central points into the most suitable area, ensuring that the points within each area maintain consistency in terms of waste collection demand intensity and road accessibility; for any given point... With the region Define the strength of affiliation: And assign the points to the areas with the highest attribution intensity: ;in, Indicates point Region The strength of belonging, and Indicates intensity weight, Indicates point With regional center The distance between them Index representing a region Indicates the first A set of points in each region. Indicates point Region The strength of belonging;

[0069] Step S34: Task matching. First, calculate the vehicle's task matching score for each region based on the vehicle's current location, remaining load, travel cost, and regional pressure value. The higher the matching score, the more suitable the vehicle is for undertaking tasks in that region. Then, perform initial matching of vehicles and regions according to the principle of maximizing matching scores. After the initial matching is completed, check capacity constraints, task uniqueness constraints, and road conflict constraints. If it is found that a vehicle is overloaded with tasks, multiple vehicles are repeatedly assigned to the same region, or there are obvious conflicts between task paths, the allocation results are readjusted through a repair mechanism until all tasks meet the executable conditions, as shown below:

[0070] ;

[0071] in, Indicates vehicle For the region Task matching score, and Indicates utility weight, Indicates the region The intensity of demand for waste removal, Indicates vehicle To the area Euclidean distance, Indicates vehicle Maximum waste carrying capacity Indicates the region The predicted total amount of garbage awaiting collection. Indicates vehicle Assign to a region , This indicates the total number of vehicles.

[0072] By performing the above operations, this solution addresses the problem that existing technologies often only consider single distance or static capacity factors in waste disposal task allocation, lacking a collaborative analysis of demand intensity, spatial correlation, and vehicle carrying capacity. This solution constructs point demand intensity and station correlation intensity, first dividing high-demand points into regions, and then matching tasks based on vehicle location, remaining load, travel cost, and regional pressure. Furthermore, it optimizes the allocation results through capacity constraints, uniqueness constraints, and conflict resolution mechanisms, making task allocation more balanced and reasonable, improving vehicle utilization, and reducing duplicate dispatches and resource waste.

[0073] Example 5, see Figure 1 and Figure 6 This embodiment is based on the above embodiment. In step S4, the cleaning and scheduling specifically includes the following steps:

[0074] Step S41: Schedule priority construction, which will be carried out in step S41. The output vehicle and area matching results are transformed into a sequence of waste collection tasks to be executed. A scheduling priority is then constructed by combining the urgency of the locations, the predicted amount of waste to be collected, and the estimated time to full load, ensuring that high-urgency, high-risk locations receive waste collection resources first. For the first... Construct scheduling priorities for each location: ;in, Indicates the first Each point at time scheduling priority, , , Indicates the priority fusion coefficient. This indicates the urgency index of waste removal. Indicates prediction up to the 1st Cumulative capacity occupied during the step Indicates the total capacity of the points. This indicates the predicted full load time; the higher the priority, the more priority that location should be served.

[0075] Step S42: Generation of transport routes. For each vehicle, under the constraints of maximum waste carrying capacity, service continuity, and return to depot, an optimal transport route is generated. Joint optimization is performed with the objectives of minimizing total travel cost, minimizing timeout risk, and maximizing priority benefits. Vehicles are defined. During the scheduling period The service path within is: ;in, and Representing vehicles The starting node and the return node, Represents the first in the path sequence One service node, Indicates vehicle Number of collection points under responsibility; Construct the scheduling objective function: ;in, This represents the comprehensive objective function of waste collection and dispatching. , , , Indicates the target weight. Indicates vehicle From node Driving towards the node Predicted travel time Indicates vehicle Maximum waste carrying capacity Indicates that the vehicle has arrived at the The estimated arrival time of each point; the following constraints must be satisfied during the optimization process:

[0076] ;

[0077] in, Indicates time vehicle Is it a service point? Decision variables, Indicates point by vehicle Inclusion path Each location is served only once by one vehicle. Indicates vehicle load constraints;

[0078] Step S43: Rescheduling. During the vehicle collection process, real-time monitoring of road traffic conditions, vehicle remaining load, and new waste growth information is conducted. When a significant change in task execution conditions is detected, a rolling re-optimization mechanism is triggered to reschedule tasks that have not yet been executed; scheduling offset is constructed: ;in, Indicates the scheduling offset. , Indicates the offset blending coefficient. Indicates point Changes in scheduling priority; when When this happens, initiate rolling rescheduling: ;in, Indicates the rescheduling trigger threshold. This indicates the remaining scheduling objective function constructed only for tasks that have not yet been executed. Indicates the updated vehicle Cleaning route sequence.

[0079] By performing the above operations, this solution addresses the problems of fixed route planning, inability to dynamically adjust to changes in road conditions and tasks, and susceptibility to overloading, timeouts, and execution conflicts in the existing construction waste removal process. It constructs scheduling priorities, generates removal routes that meet load capacity and continuous service requirements, and monitors road traffic conditions, vehicle remaining load, and changes in newly added waste in real time during execution. This triggers a rolling rescheduling mechanism to rearrange unexecuted tasks, making the removal process more real-time and adaptable, reducing the risk of overloading and delays, and improving overall removal efficiency and scheduling feasibility.

[0080] Example 6, see Figure 2 Based on the above embodiments, the present invention provides an intelligent urban construction waste collection and dispatching system based on the Internet of Things, which includes a basic construction waste dataset construction module, a waste generation demand prediction module, a task allocation module, and a collection and dispatching module.

[0081] The module for constructing a basic construction waste dataset synchronously collects multi-source information from waste generation points, road sensing devices, and vehicle terminals, and uniformly aligns data of different frequencies and formats. It also completes outlier correction and standardized coding to form a dataset that can be directly used for subsequent prediction and scheduling, and sends the data to the waste generation demand prediction module.

[0082] The waste generation demand prediction module receives data sent by the basic construction waste dataset construction module and predicts the future collection demand of each construction waste site. It first extracts the stable trend part, then separates the real demand fluctuations, and then combines multi-scale time series features and time period information to predict the future waste growth, full load time and collection urgency, obtains the demand intensity and priority service basis for each site, and sends the data to the task allocation module.

[0083] The task allocation module receives data sent by the waste generation demand prediction module and completes the allocation of collection tasks based on the prediction results. First, it integrates the demand intensity, spatial location and road conditions of each location to build a correlation. Then, it selects high-demand points as regional centers and expands them to form collection areas. Subsequently, it combines vehicle location, load capacity and regional demand to match vehicles with areas, while correcting unreasonable allocations such as conflicts and overloading, to obtain an executable task allocation and send the data to the collection scheduling module.

[0084] The waste collection and dispatch module receives data sent by the task allocation module and performs specific waste collection and dispatch for the already allocated tasks. It generates dispatch priorities based on the urgency of the location, the predicted amount of waste, and the time to full load, and then plans a waste collection route for each vehicle that meets the load and return requirements.

[0085] It should be noted that, in this document, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0086] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

[0087] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A smart urban construction waste collection and dispatching method based on the Internet of Things, characterized in that: The method includes the following steps: Step S1: Construct a basic construction waste dataset by synchronously collecting multi-source information from waste generation points, road sensing devices, and vehicle terminals, and uniformly aligning data of different frequencies and formats. At the same time, complete outlier correction and standardized coding to form a dataset that can be directly used for subsequent prediction and scheduling. Step S2: Waste generation demand forecasting. Forecast the future collection demand of each construction waste site; first extract the stable trend part, then separate the real demand fluctuation, and then combine multi-scale time series characteristics and time period information to predict the future waste growth, full load time and collection urgency, so as to obtain the demand intensity and priority service basis for each site. Step S3: Task allocation. Based on the prediction results, the task allocation for waste removal is completed. First, the demand intensity, spatial location and road conditions of each point are combined to build a correlation. Then, high demand points are selected as regional centers and expanded to form waste removal areas. Subsequently, the vehicles are matched with the areas based on vehicle location, load capacity and regional demand. At the same time, unreasonable allocations such as conflicts and overloading are corrected to obtain an executable task allocation. Step S4: Waste collection and dispatching. The specific waste collection and dispatching of the assigned tasks are carried out. The dispatching priority is generated based on the urgency of the location, the predicted amount of waste and the time of full load. Then, a waste collection route that meets the load and return requirements is planned for each vehicle.

2. The intelligent urban construction waste collection and dispatching method based on the Internet of Things according to claim 1, characterized in that: In step S2, the waste generation demand forecasting specifically includes the following steps: Step S21: Construct a trend baseline. First, extract a slowly changing trend baseline from historical waste disposal data. Then, use the difference between the current observation and the trend baseline to represent the actual demand fluctuations, separating stable growth from sudden growth. Step S22: Multi-scale temporal coding. Multi-scale convolutional coding is performed on the residual sequence, and the response at different scales is dynamically gated using temporal phase information, so that the model can capture both short-term surges caused by construction activities and slow rises caused by cycles; then attention-weighted fusion is performed on the outputs at different scales. Step S23: Predict the full load time, predict the future amount of garbage growth, and add the growth amount to the current occupied capacity to directly deduce the full load time; first, predict the future growth amount; then, recursively obtain the future capacity occupancy; finally, calculate the full load time. Step S24: Introduce constraints, integrate the growth rate, full load time, and road traffic status into a waste removal urgency index, construct a joint training loss, and add physical consistency constraints during the training phase.

3. The intelligent urban construction waste collection and dispatching method based on the Internet of Things according to claim 1, characterized in that: In step S3, the task allocation specifically includes the following steps: Step S31: Calculate the site association strength. Convert the predicted results of each construction waste site into a collection demand strength value, and construct the site association strength by combining spatial distance, road resistance and urgency. First, calculate the demand strength value of the site; then construct the association strength between any two sites. Step S32: Initial area division. Select high-demand collection points from all construction waste collection points and form an initial collection area with these high-demand collection points as the core. First, select the point with the greatest collection demand as the center point of the area. Step S33: Area expansion, incorporating non-central points into the most suitable area, so that the points within each area are consistent in terms of the intensity of waste collection demand and road accessibility; Step S34: Task matching. First, calculate the task matching score of the vehicle for each region based on the vehicle's current location, remaining load, travel cost, and regional pressure value. The higher the matching score, the more suitable the vehicle is for undertaking the task in that region. Then, perform initial matching of the vehicle and the region according to the principle of maximizing the matching score. After the initial matching is completed, check the capacity constraints, task uniqueness constraints, and road conflict constraints. If it is found that a vehicle is overloaded with tasks, multiple vehicles are repeatedly assigned to the same region, or there are obvious conflicts between task paths, the allocation results are readjusted through the repair mechanism until all tasks meet the executable conditions.

4. The intelligent urban construction waste collection and dispatching method based on the Internet of Things according to claim 1, characterized in that: In step S4, the waste collection and scheduling specifically includes the following steps: Step S41: Schedule priority construction, which will be carried out in step S41. The output vehicle and area matching results are transformed into a sequence of cleanup tasks to be executed. The scheduling priority is constructed by combining the urgency of the location, the predicted amount to be cleaned up, and the estimated full load time, so that high-urgency and high-risk locations can obtain cleanup resources first. Step S42: Generation of collection routes. For each vehicle, under the conditions of satisfying the maximum waste carrying capacity constraint, service continuity constraint, and return to the depot constraint, the optimal collection route is generated, and joint optimization is performed with the objectives of minimizing the total travel cost, minimizing the timeout risk, and maximizing the priority benefit; constraints are introduced in the optimization process. Step S43: Rescheduling. During the vehicle's collection process, the road traffic status, vehicle remaining load, and new garbage growth information are monitored in real time. When a significant change is detected in the task execution conditions, a rolling re-optimization mechanism is triggered to reschedule tasks that have not yet been executed.

5. The intelligent urban construction waste collection and dispatching method based on the Internet of Things according to claim 1, characterized in that: In step S1, constructing the basic construction waste dataset specifically includes the following steps: Step S11: Data acquisition. Data is collected synchronously from construction waste generation points, road sensing devices, and vehicle terminals. Data of different frequencies and formats are unified to the same time granularity to form the original observation vector that can be directly called for subsequent modeling. Step S12: Outlier identification, correction of outliers, drift values ​​and missing values ​​that appear during the collection process; for outliers, first construct correction weights based on neighboring sites and adjacent time points, and then replace the original outliers with spatiotemporal consistency estimates. Step S13: Standardize the coding, transform the corrected multi-source data into time window samples of a uniform scale, and explicitly add time phase information so that the model can identify daily, weekly, and construction rhythm changes.

6. An IoT-based intelligent urban construction waste collection and dispatching system, used to implement the IoT-based intelligent urban construction waste collection and dispatching method as described in any one of claims 1-5, characterized in that: It includes a module for building a basic construction waste dataset, a module for predicting waste generation demand, a task allocation module, and a waste collection and scheduling module.

7. The intelligent urban construction waste collection and dispatching system based on the Internet of Things as described in claim 6, characterized in that: The module for constructing a basic construction waste dataset synchronously collects multi-source information from waste generation points, road sensing devices, and vehicle terminals, and uniformly aligns data of different frequencies and formats. It also completes outlier correction and standardized coding to form a dataset that can be directly used for subsequent prediction and scheduling, and sends the data to the waste generation demand prediction module. The waste generation demand prediction module receives data sent by the basic construction waste dataset construction module and predicts the future collection demand of each construction waste site. First, extract the stable trend part, then separate the real demand fluctuations, and then combine multi-scale time series features and time cycle information to predict future waste growth, full load time and collection urgency, obtain the demand intensity and priority service basis for each location, and send the data to the task allocation module. The task allocation module receives data sent by the waste generation demand prediction module and completes the allocation of collection tasks based on the prediction results. First, it integrates the demand intensity, spatial location and road conditions of each location to build a correlation. Then, it selects high-demand points as regional centers and expands them to form collection areas. Subsequently, it combines vehicle location, load capacity and regional demand to match vehicles with areas, while correcting unreasonable allocations such as conflicts and overloading, to obtain an executable task allocation and send the data to the collection scheduling module. The waste collection and dispatch module receives data sent by the task allocation module and performs specific waste collection and dispatch for the already allocated tasks. It generates dispatch priorities based on the urgency of the location, the predicted amount of waste, and the time to full load, and then plans a waste collection route for each vehicle that meets the load and return requirements.

Images