Evaluation method for evaluating accumulated dust removal efficiency of road cleaning vehicle based on road underway monitoring

Abstract

The invention relates to an evaluation method for evaluating accumulated dust removal efficiency of a road cleaning vehicle based on road underway monitoring. The method belongs to the field of road cleaning, and comprises the following steps: carrying out zoning and type division on roads in a research area according to multi-dimensional factors such as road properties, traffic flow, surrounding environment and the like, selecting representative sample sections, and formulating a detailed sampling and monitoring plan; collecting road dust concentration, traffic flow and environmental parameter data at different time nodes by using an underway monitoring system in combination with fixed-point sampling; the whole operation process of the cleaning vehicle is recorded in detail; performing integration and normalization processing on the multi-source data, eliminating abnormal data and correcting environmental disturbance, and realizing comparability analysis under different conditions; based on the standardized data model, the dust retention removal efficiency and the effect duration of the cleaning vehicle under different operation and environment conditions are evaluated, and a scientific basis is provided for optimization and decision making of road cleaning operation.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of road cleaning, and more particularly relates to an evaluation method for evaluating dust removal efficiency of road cleaning vehicles based on road navigation monitoring. BACKGROUND

[0002] At present, urban road cleaning mainly relies on road cleaning vehicles for cleaning operation. The traditional dust removal effect evaluation method has the problems of strong subjectivity and inaccurate data. In China, water spraying and suction sweeping are commonly used to control road dust in cities. Different regions adopt different cleaning modes according to their own conditions. Some cities adjust the cleaning frequency according to the seasons, and some adopt a combination of multiple cleaning methods. Local governments have invested a lot of manpower, material resources and financial resources to control road dust. Therefore, how to scientifically and reasonably evaluate the effectiveness and effective duration of various road cleaning measures, so as to find out the efficient cleaning measure combination mode, has become an important research topic.

[0003] The main factors affecting the effect of road dust control measures include but are not limited to road type, traffic flow, climate conditions and other factors. In order to accurately evaluate the efficiency of different types of road dust control measures, it is necessary to quantitatively detect the changes of road dust emissions. There are two commonly used detection methods at present: AP-42 method and road resuspended aerosol dynamic emission test system (TRAKER). The AP-42 method is a standard method with low cost and simple implementation conditions. Its principle is to use a vacuum cleaner to collect road dust, and then go through a series of processing processes such as weighing and screening to calculate the dust load. Although this method has been widely used, it has the limitations of complex operation, tedious processing process, lack of safety in sampling in traffic flow, and inability to sample on wet road surface. In contrast, the TRAKER method is a vehicle-mounted road dust monitoring system based on laser sensor principle, which has the characteristics of real-time online and high efficiency and convenience, and can be used for long-time continuous monitoring in a large range, and is very suitable for efficiency evaluation research of different dust control measures. However, relying solely on the AP-42 method or the TRAKER method is still difficult to fully meet the requirements of measuring the efficiency of various road dust control measures. With the acceleration of urbanization, the quality of road cleaning directly affects the environmental quality of the city and the health of the citizens. Therefore, it is necessary to develop a more objective and accurate evaluation method to meet the needs of efficient control of road dust pollution and improve the air environmental quality.

[0004] (2) Brief description of prior art solutions

[0005] To expand the application of the TRAKER method, the AP-42 and TRAKER methods for detecting road dust were established. Efficiency assessments of road dust control measures that conform to my country's actual meteorological and climatic conditions and urban air pollution management schemes were conducted. Evaluation indicators for the effectiveness of these measures were established. The TRAKER system was used to achieve rapid, dynamic, large-scale, and high-temporal-resolution evaluation of these indicators. Factors affecting efficiency changes were analyzed to determine the most efficient combination of cleaning measures and cleaning frequency. This provided reference for local governments to formulate scientific, reasonable, economical, and effective road dust control schemes, avoiding the blind waste of large amounts of manpower, material resources, financial resources, and water resources with minimal results. (2) By establishing the connection between the TRAKER and AP-42 methods, the application scope of the TRAKER method was expanded: calculating road cleaning parameters, analyzing the spatiotemporal distribution characteristics of urban road dust; directly obtaining dust load using the TRAKER method to assess road cleanliness and the effectiveness of control measures; and conducting road pollution level evaluations, etc. Summary of the Invention

[0006] This invention constructs a set of experimental methods and evaluation index systems for assessing the dust removal efficiency of road cleaning measures. Through automated and standardized data acquisition and processing, it reduces the impact of human factors on the evaluation results, improving the objectivity and reliability of the assessment. The goal of this invention is to improve the data accuracy of dust removal effect assessment through advanced sensing technology and data processing algorithms. This invention will solve the sampling limitations in existing technologies, enabling the assessment of dust removal effectiveness under various road conditions (including wet road surfaces).

[0007] To achieve the above objectives, the present invention employs the following technical solution: the method comprises:

[0008] Road zoning and categorized sampling planning: Based on the actual conditions of urban roads, combined with road characteristics, traffic flow, surrounding environment and historical dust load records, the roads in the study area are divided into zones and types, and representative road sections are selected to formulate a sampling and monitoring plan to ensure the comprehensiveness and comparability of the data;

[0009] After clarifying the boundaries and scope of the study area, the roads in the area are systematically sorted and grouped based on multiple dimensions such as road traffic function level, pavement material, dynamic traffic flow, sensitive attributes of the environment on both sides and around the road, and historical air quality monitoring data.

[0010] Based on the above multi-factor analysis, spatial clustering was performed according to the similarity of key dimensions, dividing the study area into several assessment zones with homogeneous traffic and environmental characteristics. Within each zone and between zones, road type groups were further subdivided according to the main road types. Taking into account the representativeness, accessibility and safety of spatial distribution, representative sample segments were selected for each road type group.

[0011] On the selected sample sections, a mobile monitoring system was used for continuous monitoring and data collection. By combining high-frequency mobile monitoring with fixed-point sampling, road dust data before and after cleaning and at different time points were obtained, and relevant environmental parameters, traffic flow information and meteorological data were recorded.

[0012] The operation process of road cleaning vehicles is recorded simultaneously, including operation time, operation mode, operation intensity, equipment parameters, etc., and the recorded results are correlated with mobile monitoring data;

[0013] The collected road dust data, traffic data, environmental parameters, cleaning operation records, etc. are integrated and normalized from multiple sources to establish a standardized data analysis model, eliminate extreme data interference, and unify the data caliber under different external conditions.

[0014] Based on standardized data, this study analyzes the efficiency of road cleaning vehicles in removing road dust under different operating conditions, evaluates their short-term dust removal rate, duration of cleaning effect, and optimization effect of operating strategies, and provides a scientific basis for road cleaning decisions.

[0015] In one approach, the road zoning and categorized sampling planning steps further include: using a systematic approach, classifying all roads within the study area in detail based on their traffic function level, pavement material, and traffic flow data. Main roads, secondary roads, and branch roads are processed separately according to their traffic load and service functions. Spatial attributes are further subdivided using geographic information system analysis tools to ensure the systematicity and representativeness of the sample segments and reduce sampling bias.

[0016] In one approach, road zoning and classification are further integrated with the environmental sensitivity attributes of both sides and the surrounding area, including the distribution of construction sites, concentrated factory areas, schools and hospitals, as well as the coverage of commercial areas and green belts. By comprehensively quantifying these environmental characteristics, the environmental impact types of roads are refined, and zoning based on environmental exposure risk is achieved.

[0017] In one approach, after road zoning is completed, the traffic flow of representative road segments is dynamically assessed using historical traffic flow monitoring data and existing intelligent transportation facilities. This ensures that the selected segments are typical in terms of traffic flow distribution, and that the segment length is no less than 500 meters to effectively buffer the impact of traffic fluctuations on the monitoring results. The geometric characteristics and traffic organization methods of the segments are also taken into account.

[0018] In one approach, the mobile monitoring system collects road dust concentration data on selected sample sections in a high-frequency, continuous manner, and simultaneously collects meteorological parameters and traffic flow information. The collection process should cover multiple time points before, after, and after cleaning operations. Through high-resolution spatial and temporal data, a spatiotemporal distribution sequence reflecting changes in road dust is formed.

[0019] In one approach, the records of the cleaning vehicle's operation process include, but are not limited to, operation time, operation path, operation speed, water spray volume, brushing speed, and type of dust suppressant used. All operation parameters are stored in the form of a structured log and matched with dynamic monitoring data of road dust accumulation to achieve full-process traceability and data-driven management of the operation behavior.

[0020] In one approach, during the data integration and normalization process, multivariate normalization and outlier removal strategies are employed to address environmental disturbances such as wind speed, humidity, and road humidity, as well as traffic flow fluctuations. This ensures the comparability of sampling data under different external environmental conditions and enhances the universality and accuracy of the evaluation method through standardized data models.

[0021] In one approach, data analysis results include not only an assessment of the instantaneous dust removal rate of a single cleaning operation, but also the duration of the cleaning effect and the trend of efficiency changes under different cleaning modes. By comparing and analyzing the impact of road type, environmental attributes, and operational parameters on removal efficiency, dynamic assessment of dust removal efficiency and intelligent optimization of operational strategies in complex urban road environments can be achieved, providing support for smart sanitation operation decisions.

[0022] In one approach, the final assessment results can be output in multi-dimensional, hierarchical reports or visualizations, such as by zoning, road type, and operation mode. This supports dynamic updates and knowledge accumulation, forming a sustainable assessment knowledge base for urban road cleaning operations, and providing a data foundation for policy formulation and management optimization.

[0023] Beneficial effects of this invention:

[0024] 1. This invention reduces the impact of human factors on the evaluation results and improves the objectivity and reliability of the evaluation by automating and standardizing the data collection and processing process.

[0025] 2. This invention employs advanced sensor technology and data processing algorithms, which improves the accuracy of data for evaluating the dust removal effect.

[0026] 3. This invention simplifies the operation process for evaluating the dust removal effect, reduces the difficulty of operation, and reduces the possibility of errors.

[0027] 4. This invention solves the problem of sampling limitations in the prior art, enabling the evaluation method to assess the dust removal effect under various road conditions (including wet road surfaces).

[0028] 5. This invention provides a method for long-term, large-scale continuous monitoring, which improves the comprehensiveness and timeliness of the assessment.

[0029] 6. This invention significantly improves the efficiency of dust removal effect evaluation through automated monitoring and data analysis. Attached Figure Description

[0030] Figure 1 This is a flowchart of the method of the present invention;

[0031] Figure 2 A schematic diagram of the TRAKER equipment system installation. Detailed Implementation

[0032] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0033] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. To facilitate understanding, the invention will now be described more fully with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to make the disclosure of the invention more thorough and complete.

[0034] like Figure 1 As shown, an evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring includes the following steps:

[0035] Step 1: Road Zoning and Classification Sampling Planning. Based on the actual conditions of urban roads, combined with road type (main roads, secondary roads, branch roads, etc.), traffic flow, surrounding environment, and historical dust records, the roads in the study area are divided into zones and types. Representative road sections are selected, and a detailed sampling and monitoring plan is developed to ensure the comprehensiveness and comparability of the data.

[0036] The implementation process of road zoning and categorized sampling planning first requires clarifying the boundaries and scope of the study area, and then systematically sorting out all roads within the area based on multi-dimensional factors. The core basis is the classification of basic road properties, such as traffic function level (main roads carry cross-regional traffic flow, secondary roads undertake regional distribution functions, and branch roads serve community microcirculation) and pavement materials (differences such as asphalt and cement). At the same time, it deeply integrates dynamic traffic flow data (estimate daily / peak hour traffic flow based on data from intelligent traffic checkpoints or historical traffic monitoring stations), environmental sensitivity attributes of both sides and surrounding areas of the road (such as the distribution of construction sites, concentrated factory areas, sensitive points such as schools and hospitals, commercial areas or green belt coverage), and information on road sections with high dust loads recorded by historical air quality monitoring stations or previous studies.

[0037] After completing the multi-factor coupling analysis described above, spatial clustering is required based on the similarity of roads in key dimensions (e.g., high-traffic arterial roads adjacent to construction sites) to divide the study area into several evaluation zones with relatively homogeneous environmental and traffic characteristics. Within each zone and between zones, road type groups with evaluation significance are further subdivided according to the main road type (e.g., main logistics channels with a high proportion of heavy trucks, side roads adjacent to restaurant streets, etc.). Under this hierarchical framework, specific sample segments are carefully selected for each road type group, taking into account the representativeness, accessibility, and safety of the spatial distribution. During the selection process, it is necessary to ensure that the sample segment length is sufficient (usually ≥500 meters to buffer the impact of traffic fluctuations), and that the geometric features (e.g., slope, curves) and traffic organization methods (e.g., whether there is separation of motor vehicles and non-motorized vehicles) have the general characteristics of that type. Finally, a detailed plan was developed for the schedule, frequency, and triggering conditions (such as specific wind speed ranges) of basic mobile monitoring (including baseline and post-operation tracking) and auxiliary point sampling (AP-42 method) for each sample section. Traffic flow control requirements during monitoring were clearly defined (such as avoiding major event periods or periods of extreme congestion). Standard operating procedures for recording instantaneous environmental changes in the surrounding area (such as earthwork operations at nearby construction sites or interference from water trucks) were also developed to ensure the spatial coverage breadth, road type coverage depth, and strict comparability of the core data obtained across sample sections and periods.

[0038] like Figure 2As shown, Step 2: Multi-source data baseline acquisition. Using the TRAKER vehicle-mounted mobile monitoring system, baseline monitoring of the target road section was conducted under multiple time periods and weather conditions before road sweeping operations. This obtained the spatial distribution of particulate matter concentrations such as PM10 and PM2.5. Simultaneously, environmental parameters such as wind speed and humidity were collected from meteorological sensors to form a multi-dimensional database of road dust conditions. To compensate for the shortcomings of the TRAKER method in particulate matter classification and quantitative analysis of surface dust load, a small amount of AP-42 sampling was used to quantitatively calibrate the dust load for different particle size ranges.

[0039] The implementation of multi-source data baseline acquisition begins with comprehensive pre-operational environmental baseline monitoring of the target road sections. At least one week before the start of road cleaning operations, vehicles equipped with the TRAKER mobile monitoring system repeatedly travel along pre-planned sampling routes on the target road sections, covering different time periods on typical days (focusing on morning and evening traffic peaks, midday off-peak periods, and low-traffic periods at night) and varying meteorological conditions (such as sunny and dry days, wet days after rain, and periods with different wind speeds). During monitoring, the TRAKER system collects high spatial resolution PM2.5 and PM10 concentration data in real time using optical scattering methods. Simultaneously, onboard meteorological sensors continuously record second-by-second wind speed, wind direction, relative humidity, and temperature parameters. All spatiotemporal data are precisely matched to road location coordinates using GPS positioning information, forming an initial five-dimensional dataset of "time-space-pollutant concentration-meteorological elements". To overcome the limitations of the TRAKER method in quantifying particle size distribution and absolute physical load, surface dust samples were collected at representative points along each target road segment (such as the road centerline and both side curbs) according to the AP-42 standard. A stiff brush and a square mold (typically 30cm × 30cm) were used to precisely collect inhalable dust samples (<100μm). The samples were then weighed using laboratory grading sieves to obtain the particle size distribution data (unit: g / m³) for PM2.5 and PM10. 2 The physical measurement results at this point scale were regressed with the instantaneous concentration data from the TRAKER patrol at the same location to construct a concentration-load conversion function (such as a calibration coefficient matrix) for each particle size segment. Then, spatial interpolation technology was used to extrapolate the point calibration results to the entire road segment, ultimately generating a comprehensive road dust database that includes both dynamic particulate matter concentration distribution and quantitative physical load indicators, thus establishing a high-precision comparison benchmark for subsequent cleaning effect evaluation.

[0040] Step 3: Synchronously record the entire road cleaning operation process. Combining the Internet of Things (IoT) and vehicle positioning systems, record the start and end times, operation methods (watering, vacuuming, sweeping, or a combination), speed, and routes of road cleaning vehicle operations throughout the entire process, achieving digitization and standardization of operation information. When necessary, collect parameters such as vehicle configuration and cleaning agent type to supplement operational details affecting cleaning efficiency.

[0041] Based on the completion of road zoning and baseline data collection, the entire road cleaning operation process is digitally recorded through a technical system deeply integrated with the previous work. According to the representative road sections defined in Step 1 and the established sampling plan, cleaning vehicles entering the target road sections are equipped with IoT terminal modules (integrating high-precision GPS / BeiDou positioning, 4G / 5G transmission, and multi-channel sensor interfaces). During operation, the positioning system transmits the vehicle's real-time latitude and longitude coordinates at a rate of seconds. By automatically matching with the road section GIS layer, it dynamically records whether the operation path strictly covers the planned sample section's spatial range. Simultaneously, the vehicle control bus captures the operation mode signal status through a digital interface (distinguishing between simple mechanical sweeping, pure water spraying for dust suppression, or combined sweeping and flushing operations), and obtains the real-time operation speed through an integrated vehicle speed sensor or CAN bus. All spatiotemporal parameters (location, speed, operation mode) and status timing signals (equipment start / stop timestamps, water pump / fan on / off status) are transmitted back to the cloud platform in real time via the IoT network, automatically generating a structured four-dimensional digital operation log of "road-time-operation mode-speed".

[0042] To refine the analysis of differences in operational intensity, the system simultaneously collects auxiliary operating parameters when necessary: ​​for water sprinkler vehicles, it records the water volume sprayed per unit distance (L / km) via a flow meter and the nozzle pressure (MPa) via a pressure sensor; for sweeper vehicles, it records the roller brush speed (rpm) and the dust collection box negative pressure value (kPa); if environmentally friendly dust suppressants or special cleaning agents are used, the agent's component type, dilution ratio, and application amount are linked to the operation record via vehicle-mounted barcode scanning or manual input. These parameters are automatically categorized and stored according to the road type defined in step 1 (e.g., high-pressure washing is required for heavy-load sections of main roads), and a unique association code is established with the corresponding road section dust baseline data established in step 2. Finally, through automated verification and standardized cleaning on the cloud platform (e.g., removing ultra-low-speed anomalies caused by traffic congestion), a complete, traceable, and parameter-structured "electronic operation file" is formed for each road sample section during the cleaning process, providing core operating parameter support for subsequent accurate benchmarking of different dust backgrounds and effectiveness verification under different operation modes.

[0043] Step 4: Continuous mobile monitoring after cleaning operations. Immediately after the cleaning operation is completed, the TRAKER system is activated to conduct repeated mobile monitoring of the same road segment at multiple time points (such as 0h, 2h, 4h, 8h, 24h, and 48h) to obtain the dust concentration recovery curve, thereby realizing dynamic tracking of the cleaning removal efficiency and the duration of the effect. At the same time, environmental meteorological parameters and traffic flow information are continuously collected synchronously.

[0044] Continuous mobile monitoring after road cleaning operations are completed aims to dynamically capture the recovery pattern of dust concentration. Its implementation requires close coordination with the operation recording system. As soon as the cleaning vehicle leaves the end of the target road section, the pre-set monitoring plan is immediately activated: the monitoring vehicle equipped with the TRAKER system performs the first operation along the same trajectory as in step 2 and then immediately performs mobile monitoring (marked as the 0h baseline). Subsequently, the same road sample section is completely repeated with mobile monitoring according to the preset tracking cycle (such as key time nodes such as 2h, 4h, 8h, 24h, 48h, etc.). During each monitoring session, the TRAKER system not only reproduces the particulate matter concentration indicators (PM10, PM2.5) collected at the initial baseline, but also simultaneously activates the vehicle-mounted weather station to collect real-time wind speed, wind direction, temperature, humidity, and precipitation status (identifying micro-precipitation events through a rain sensor). At the same time, it extracts the traffic flow, vehicle type composition (especially the proportion of heavy vehicles), and average vehicle speed within the corresponding monitoring window period through roadside fixed traffic flow monitoring equipment or floating car GPS data. All time-series data streams are matched to the meter-level segmented units of the road chain through spatial registration technology.

[0045] To achieve high-precision concentration recovery curve construction, the monitoring process strictly adheres to the "three same" principle: the same road segment (ensuring spatial comparability), the same navigation route (eliminating differences in concentration gradients across road cross sections), and the same instrument parameter configuration (maintaining consistency in optical sensor scattering coefficients, etc.). For key dust-sensitive locations (such as parking lines at intersections, uphill sections, or near construction site exits), low-speed repeated traversal (speed deviation controlled within ±5 km / h) is added to the regular navigation to obtain the average concentration of local hotspots. The massive time-series data stream generated by the monitoring is synchronously input into the cloud processing platform. After automatic verification and removal of instrument anomalies, it is used for spatiotemporal correlation calculations with the operational mode parameters recorded in step 3: for example, quantifying the coupling relationship between the dust rebound rate 2 hours after high-pressure washing and the traffic flow increase section, or analyzing whether the dust concentration 24 hours later is disturbed by nighttime high-wind events without precipitation and returns to the pre-operation level. Ultimately, a three-dimensional dynamic evolution dataset of "dust concentration-meteorology-traffic" based on timestamps was formed for each sample road, providing empirical support for the subsequent establishment of empirical models of cleaning efficiency and effect half-life prediction under different road types and operation modes.

[0046] Step 5: Multi-factor data integration and denoising. Based on the collected multi-time period and multi-source data, irrelevant or abnormal data (such as periods affected by sudden rainfall or extreme traffic events) are removed using algorithms. Normalization correction is performed using environmental meteorological parameters and traffic intensity to ensure effective data comparability and interpretability.

[0047] Step 5, multi-factor data integration and denoising, achieves deep fusion of multi-source heterogeneous data by constructing a unified spatiotemporal data cube. In the data integration phase, the road dust baseline data collected in Step 2 (including AP-42 physical load calibration parameters), the operational parameters recorded in Step 3 (timestamp-location-mode-intensity four-dimensional tensor), the concentration time-series recovery curve monitored in Step 4, and the meteorological and traffic time-series monitoring data are linked by primary key using road chain spatial coding (e.g., GIS road network topology ID) and UTC timestamps to form a high-dimensional data array (e.g., [road segment × time point × monitoring level × variable type]). Figure 1 The mathematical expression shown is as follows:

[0048] D(i,t,k,l)∈R N ×T×K×L

[0049] Where i represents the road segment index, t represents the time point, k represents different monitoring levels, and l represents the variable dimension.

[0050] Outlier removal employs a multi-layered cascaded algorithm: First, box plot rules (mathematical formula: Q1-1.5IQR>data>Q3+1.5IQR are used to mark outliers, where IQR=Q3-Q1) to eliminate extreme values ​​caused by sensor hardware failures; second, meteorological threshold judgments (e.g., activating the "rain interference flag" when precipitation > 0.5mm / h) are combined to discard data slices contaminated by sudden meteorological events for the entire time period; finally, dynamic spatiotemporal clustering (based on the DBSCAN algorithm, distance function) is used.

[0051] dist=α·|ΔPM 10 |+β·|Δwind speed|+γ·|Δtraffic flow|)

[0052] This feature identifies abnormal events such as traffic congestion or sudden emissions from construction sites. The percentage of valid data retained after cleaning must reach at least 85% of the total data collected.

[0053] Multi-factor normalization correction is achieved by constructing an environmental background response model: defining the change in dust concentration.

[0054] ΔC(i,t)=C post (i,t)-C pre (i)

[0055] (C pre (i) Taking the baseline mean from step 2, establish the multivariate linear compensation model as shown in formula (1):

[0056]

[0057] Where X mThe standardized value of the m-th disturbance factor (including logarithmic transformation of wind speed ln(v+1), linear term of relative humidity RH, and heavy vehicle equivalent Q) HGV (etc.), coefficient β m LASSO regression (loss function:) was performed using time-series data from the entire road segment. The results are estimated. For road segments with periodic fluctuations (such as school zones), a seasonal decomposition algorithm (STL: C(t) = Trend(t) + Seasonal(t) + Residual(t)) is further introduced to eliminate the influence of intraday traffic patterns.

[0058] Finally, the standardized cleaning effect index defined by formula (2) is generated:

[0059]

[0060] Here, τ represents a specific time window after the task (e.g., the 0-2h effect period), W mode The weighting factor for the work mode (generated by regression of parameters such as water pressure and vehicle speed in step 3), and δ, the time decay coefficient, serve as the dimensionless evaluation benchmark for subsequently quantifying the core effectiveness of different work modes. The processed database needs to undergo cross-validation to ensure RMSE ≤ 15% and R0. 2 The model explanatory power is ≥0.8.

[0061] Step 6: Comprehensive Evaluation Modeling of Efficiency and Duration. A comprehensive evaluation model of "instantaneous removal efficiency - duration of effectiveness" is constructed. This model uses the reduction in particulate matter concentration after cleaning (removal rate) and the duration of low dust load as dual indicators. Combined with characteristic parameters such as road type, cleaning method, and work intensity, regression or machine learning methods are used to quantify the actual effects of various cleaning modes. This clarifies the optimal dust removal efficiency and its duration under different combinations, achieving a differentiated and scientific evaluation of road cleaning measures.

[0062] After completing the fusion and normalization of multi-source data, step 6 achieves the scientific quantification of cleaning efficiency by constructing a dual-response variable regression model. The modeling basis is the standardized database generated in step 5, and the core response indicator is defined as: instantaneous removal efficiency.

[0063] η max (i)=[1-min(ΔC corrected (i,t∈[0h,0.5h]) / C pre (i)]×100%

[0064] (Characterizing the maximum dust removal rate achieved within 30 minutes after the operation), duration

[0065] T sustain (i) = argmax t {Cpost (i,t)≤0.4×C pre (i)}

[0066] (i.e., the duration for which dust concentration remains below 40% of the baseline value, in hours). Classify road types (as per step 1, X). type ∈{1,2,...,6}), Operation method (X) mode One-hot encoding is used to represent watering / vacuuming / combined operation), and the work intensity parameter (X). intensity The 10-dimensional feature vector, including the average water pressure P, the roller speed ω, and the equivalent operating speed v collected in step 3, is used as explanatory variables.

[0067] The model structure adopts a hierarchical estimation framework: first, η is established. max Relationship with eigenvectors:

[0068]

[0069] Where the function f k (·) The nonlinear interaction effect of the operation method is fitted by a feedforward neural network (hidden layer activation function ReLU, loss function Huber loss), v std The standard design speed is used; secondly, a Weibull survival model is constructed for the duration.

[0070]

[0071] In the risk function h(t), the shape parameter ρ controls the baseline risk curve, and the accelerated failure factor exp(·) includes an interaction term between road class and operating parameters. The joint optimization of the parameters of the two models adopts regularized maximum likelihood estimation (elastic network penalty term: 0.3‖α‖1+0.7‖α‖2).

[0072] To analyze the optimal combination of task parameters, a three-dimensional decision space is constructed based on a dual response surface model: given road type constraints (e.g., main road X) type =1), solve the objective function max[ω η ·η max +ω T ·T sustain The weighting factor ω is determined by the AHP (Analytic Hierarchy Process), and the decision variables are (P, ω, v) ∈ the feasible region. A contour map of the cleaning strategy is generated using a Gaussian process surrogate model.

[0073] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0074] It should be understood that the above detailed description of the technical solutions of the present invention with reference to preferred embodiments is illustrative and not restrictive. Those skilled in the art can modify the technical solutions described in the embodiments or make equivalent substitutions for some of the technical features based on reading this specification; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating the dust removal efficiency of road cleaning vehicles based on road mobile monitoring, characterized in that, The method includes: Road zoning and categorized sampling planning: Based on the actual conditions of urban roads, combined with road nature, traffic flow, surrounding environment and historical dust records, the roads in the study area are divided into zones and types, and representative road sections are selected. After clarifying the boundaries and scope of the study area, the roads in the area are systematically sorted and grouped based on multiple factors such as road traffic function level, pavement material, dynamic traffic flow, sensitive attributes of the environment on both sides and around the road, and historical air quality monitoring data. Based on multi-factor analysis, spatial clustering was performed according to the similarity of key dimensions, dividing the study area into several assessment zones with homogeneous traffic and environmental characteristics. Within each zone and between zones, road type groups were further subdivided according to the main road types. On the selected sample sections, a mobile monitoring system was used for continuous monitoring and data collection. By combining high-frequency mobile monitoring with fixed-point sampling, road dust data before and after cleaning and at different time points were obtained, and relevant environmental parameters, traffic flow information and meteorological data were recorded. The operation process of road cleaning vehicles is recorded simultaneously, including operation time, operation mode, operation intensity, and equipment parameters, and the recorded results are correlated with mobile monitoring data; The collected road dust data, traffic data, environmental parameters, and cleaning operation records are integrated and normalized from multiple sources to establish a standardized data analysis model, eliminate extreme data interference, and unify the data caliber under different external conditions. Based on standardized data, this study analyzes the efficiency of road cleaning vehicles in removing road dust under different operating conditions, evaluates their short-term dust removal rate, duration of cleaning effect, and optimization effect of operating strategies, and provides a scientific basis for road cleaning decisions.

2. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that: The road zoning and categorization sampling planning steps further include: using a systematic approach, classifying all roads within the study area in detail based on their traffic function level, pavement materials, and traffic flow data.

3. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, When classifying and categorizing roads, we further consider the environmental sensitivity attributes of both sides and the surrounding area, including the distribution of construction sites, concentrated factory areas, sensitive points such as schools and hospitals, as well as the coverage of commercial areas and green belts.

4. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, After the road zoning is completed, the traffic flow of representative road sections is dynamically evaluated using historical traffic flow monitoring data and existing intelligent transportation facilities to ensure that the selected sections are typical in terms of traffic flow distribution. At the same time, the length of the sections is not less than 500 meters to buffer the impact of traffic fluctuations on the monitoring results, and the geometric characteristics and traffic organization methods of the sections are also taken into account.

5. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, The mobile monitoring system collects road dust concentration data on selected sample sections in a high-frequency continuous manner, and simultaneously collects meteorological parameters and traffic flow information. The collection process should cover multiple time points before, after, and after cleaning. Through high-resolution spatial and temporal data, a spatiotemporal distribution sequence reflecting changes in road dust is formed.

6. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, The records of the cleaning vehicle's operation process include, but are not limited to, operation time, operation route, operation speed, water spray volume, brushing speed and the type of chemicals used. All operation parameters are stored in the form of structured logs and matched with dynamic monitoring data of road dust accumulation to achieve full-process traceability and data-driven management of operation behavior.

7. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, During the data integration and normalization process, multivariate normalization and outlier removal strategies are adopted to address environmental disturbances such as wind speed, humidity, road humidity, and traffic flow fluctuations, ensuring the comparability of sampled data under different external environmental conditions.

8. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, The data analysis results include not only the assessment of the instantaneous dust removal rate of a single cleaning operation, but also the duration of the cleaning effect and the trend of efficiency changes under different cleaning modes. The impact of road type, environmental attributes, and operation parameters on removal efficiency is analyzed by comparison.

9. The evaluation method for assessing the dust removal efficiency of road cleaning vehicles based on road mobile monitoring as described in claim 1, characterized in that, The final assessment results can be output in the form of multi-dimensional, hierarchical reports or visualizations based on zoning, road type, and operation mode, supporting dynamic updates and knowledge accumulation, and forming a sustainable assessment knowledge base for urban road cleaning operations.

Images