Highway fog elimination method and system based on air cannon principle

Abstract

The application discloses a highway fog elimination method and system based on the principle of air cannon, and belongs to the technical field of air cannon control, which comprises the following steps: acquiring multi-source environment and real-time traffic flow data of a target road section; extracting the visibility change trend and the wind field space-time distribution characteristics to construct an environment dynamic characteristic vector; establishing an environment energy field assistance model to calculate a coupling gain coefficient, and combining historical spraying events for optimization; dynamically generating air cannon cooperative control instruction sets, generating executable driving instructions to trigger spraying after system safety boundary verification and mechanism adaptability verification; and updating data to evaluate the effect and feed back the model for correction after spraying. The multivariate energy field model of the coupled traffic environment and the parameter self-learning correction strategy based on closed-loop feedback are adopted, so that the optimal cooperative control of the air cannon array spraying timing, angle and intensity can be realized.

Description

Technical Field

[0001] This invention relates to the field of air cannon control technology, and in particular to a method and system for eliminating fog on highways based on the principle of air cannons. Background Technology

[0002] Currently, in the field of highway safety management, the active elimination technology for sudden meteorological disasters such as fog patches is the fog patch elimination system based on the principle of air cannons. It releases high-pressure gas instantaneously to disturb the local atmosphere, causing fog droplets to spread or settle, thus achieving rapid intervention at the road section level.

[0003] In related technologies, Chinese invention patent CN115437433A discloses a remote control system for an on-site fog cannon with environmental factor control, including a main control system; an environmental monitoring system for monitoring environmental parameters around the fog cannon and connected to the main control system; and a fog cannon control system electrically connected to the main control system. The fog cannon control system includes at least a controller and a servo actuator electrically connected to the controller. The controller is connected to the main control system to receive instructions from the main control system and transmit these instructions to the servo actuator, which drives the fog cannon to change its spray angle and spray direction.

[0004] However, existing solutions of this kind still have significant shortcomings when applied to highway fog dissipation scenarios. The control closed loop heavily relies on remote manual decision-making and command issuance, resulting in a sluggish overall response due to the rapid formation and short dissipation window of fog, making it difficult to meet real-time intervention requirements. The control logic is mainly based on comparisons between instantaneous parameters at monitoring points and simple thresholds, resulting in a single decision-making dimension. It fails to fully consider the profound impact of complex environmental energy, such as dynamic changes in the wind field and traffic disturbances, on the intervention effect, leading to intervention actions that may contradict natural environmental dynamics, resulting in poor efficiency and high energy consumption. The system design does not emphasize the data acquisition anti-interference capabilities, circuit protection, and the reliability of coordinated control of actuators required for long-term operation under harsh outdoor conditions, affecting the overall stability and availability of the system. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a method and system for eliminating fog on highways based on the principle of air cannons. By constructing a multivariable energy field model coupled with the traffic flow environment and employing a parameter self-learning correction strategy based on closed-loop feedback, it is possible to achieve optimal coordinated control of the spray timing, angle, and intensity of the air cannon array.

[0006] The above objectives can be achieved through the following approach:

[0007] S1. A method for eliminating fog on highways based on the principle of air cannons, including acquiring multi-source environmental data and real-time traffic flow data of the target road section and generating a set of original monitoring data;

[0008] S2. Process the original monitoring data set, extract the visibility change trend features and the spatiotemporal distribution features of the wind field, and generate an environmental dynamic feature vector.

[0009] S3. Based on the environmental dynamic feature vector and the real-time traffic data, construct an environmental energy field-assisted model to calculate the coupling gain coefficient of the expected trajectory of the air cannon, and optimize it based on a preset historical spray event sequence.

[0010] S4. Based on the optimized coupling gain coefficient and the visibility change trend characteristics, dynamically generate the collaborative control instruction set for the air gun;

[0011] S5. Perform system security boundary verification and actuator adaptability verification on the cooperative control instruction set, and generate executable drive instructions;

[0012] S6. Based on the executable drive command, drive the telescopic motor and the electronically controlled valve to perform coordinated actions to trigger the air cannon to spray;

[0013] S7. After the air cannon is sprayed, the original monitoring data set is updated to evaluate the fog elimination effect and fed back to the environmental energy field assist model for correction.

[0014] Based on the same inventive concept, this invention also provides a highway fog elimination system based on the principle of an air cannon. The system includes: an environmental traffic flow sensing module for acquiring multi-source environmental data and real-time traffic flow data of the target road section, generating a raw monitoring data set; a feature extraction module for processing the raw monitoring data set, extracting visibility change trend features and wind field spatiotemporal distribution features, generating an environmental dynamic feature vector; and a coupling analysis module for constructing an environmental energy field-assisted model based on the environmental dynamic feature vector and the real-time traffic flow data, calculating the coupling gain coefficient of the expected trajectory of the air cannon, and based on preset historical spray data. The event sequence is optimized; an intelligent decision-making module is used to dynamically generate a set of coordinated control instructions for the air cannon based on the optimized coupling gain coefficient and the visibility change trend characteristics; an instruction generation module is used to perform system safety boundary verification and actuator adaptability verification on the coordinated control instruction set to generate executable drive instructions; an instruction execution module is used to drive the telescopic motor and the electronically controlled valve to perform coordinated actions based on the executable drive instructions to trigger the air cannon to spray; and a feedback correction module is used to update the original monitoring data set after the air cannon sprays to evaluate the fog elimination effect and feed it back to the environmental energy field assistance model for correction.

[0015] Compared with the prior art, the present invention has the following advantages:

[0016] This invention, by exploring the coupling effect of the spatiotemporal distribution of wind field and the propagation of air disturbance by real-time traffic flow, can calculate the optimal trajectory and energy dissipation of the air cannon spray in a complex dynamic flow field. This improves the targeting of fog elimination while using natural wind field and traffic-induced wind field assistance to reduce the overall energy consumption of the system and expand the effective operating coverage area.

[0017] After generating the collaborative control instruction set, this invention performs boundary checks on the current physical state of the system and verifies mechanical characteristics such as the achievability of motor torque. It then smooths or corrects high-frequency, high-load actions, avoiding hardware damage or mechanical failures caused by extreme calculation instructions. This ensures long-term stable operation and high-fidelity action response of field equipment under harsh weather conditions.

[0018] This invention assesses visibility improvement and fog dissipation area immediately after spraying and feeds these performance indicators back to the model. The system can continuously correct model parameters using updated historical spraying event sequences. As the operating time progresses, it can self-evolve according to the unique meteorological evolution patterns of road sections, continuously improving the efficiency and intelligence of handling sudden fog patches. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating a highway fog elimination method based on the principle of an air cannon, according to an embodiment of the present invention.

[0021] Figure 2 This is a timing diagram illustrating the control of visibility changes and coupling gain coefficients according to an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of the structure of a highway fog elimination system based on the principle of an air cannon, according to an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Reference Figure 1 One embodiment of the present invention proposes a method for eliminating fog on highways based on the principle of air cannons. By constructing a multivariable energy field model coupled with the traffic flow environment and a parameter self-learning correction strategy based on closed-loop feedback, the optimal coordinated control of the spray timing, angle and intensity of the air cannon array can be achieved.

[0025] The method described in this embodiment specifically includes:

[0026] S1. Acquire multi-source environmental data and real-time traffic flow data of the target road segment to generate a raw monitoring data set;

[0027] Optionally, the step of acquiring multi-source environmental data and real-time traffic flow data of the target road segment to generate an original monitoring data set includes:

[0028] Acquire temperature and humidity data and atmospheric pressure data collected by meteorological sensors installed along the target road section, and real-time visibility data collected by visibility meters, to generate a basic environmental monitoring data set;

[0029] Acquire horizontal wind speed, vertical wind speed, and wind direction data collected by a wind field detection device set above the target road section, and generate a three-dimensional wind field data set;

[0030] The system acquires vehicle speed, vehicle density, and vehicle type data collected by vehicle detectors within the target road segment, and generates a real-time traffic flow data set.

[0031] The basic environmental monitoring data set, the three-dimensional wind field data set, and the real-time traffic flow data set are aligned and fused according to a unified timestamp and spatial coordinates to generate an original monitoring data set with spatiotemporal labels.

[0032] Specifically, the step of acquiring multi-source environmental data and real-time traffic flow data of the target road section to generate the original monitoring data set involves collecting data through sensors deployed in physical space and then performing spatiotemporal alignment and fusion processing to form a structured data set. Meteorological sensors are installed on guardrails or pillars along both sides of the highway, spaced at 200-meter intervals, to measure air temperature, relative humidity, and atmospheric pressure. Visibility meters are installed at the same location as the meteorological sensors, measuring real-time visibility based on the forward scattering principle. These devices sample at a frequency of 1 Hz, with temperature measured in degrees Celsius, humidity in percentage, pressure in hectopascals, and visibility in meters, collectively forming the basic environmental monitoring data set. Wind field detection devices, employing Doppler sonar or lidar, are installed on gantry cranes or high poles above the road section, covering a range from the road surface to 10 meters in the air, measuring horizontal wind speed, vertical wind speed, and wind direction angle. Horizontal and vertical wind speeds are measured in meters per second, and wind direction angles in degrees. This device samples at a frequency of 2 Hz to generate a three-dimensional wind field data set. Vehicle detectors, employing microwave radar or video recognition equipment, are deployed along road sections to collect data on vehicle speed, vehicle density, and vehicle type classification based on vehicle length at monitored sections. Speed ​​is measured in meters per second, density in vehicles per kilometer, and vehicle types are categorized into small, medium, and large vehicles, each represented by a code. The sampling frequency is 5 Hz, generating a real-time traffic flow dataset. To integrate this heterogeneous data, a unified clock source is used for time synchronization across all devices, with timestamps... The accuracy reaches the millisecond level. Each sensor is assigned a unique geospatial coordinate. ,in and For local plane coordinates established based on the starting point of the road segment, Here are the altitude coordinates, all in meters. The alignment and fusion process is described as follows: for each synchronization time window... Extract data tuples from each monitoring point in the basic environmental monitoring dataset. ,in Represents temperature. Represents humidity. Represents pressure, Represents visibility, subscript Represents a timestamp. Represents the spatial coordinate index of the device. Extracts the data tuple for each detection node from the 3D wind field dataset. ,in Represents horizontal wind speed. Represents vertical wind speed. Represents wind direction angle, subscript Represents the coordinate index of the wind farm equipment. Extracts data tuples from each detection section of the real-time traffic flow dataset. ,in Represents vehicle speed. Represents vehicle density. Represents vehicle model code, subscript Represents the coordinate index of the vehicle detector. Based on the timestamp. Align the above data to the same time point, and based on spatial coordinates. The nearest spatial monitoring data are correlated. After fusion, a raw monitoring data set with spatiotemporal labels is generated. Each of its data entries is represented as This entry contains complete environmental and traffic flow status information at a specific time and spatial location. Time synchronization accuracy and spatial coordinate calibration are based on GPS and road segment mapping data.

[0033] For example, step S1 is implemented for the target road section from K100+000 to K100+500 of the G50 expressway. Weather sensors and visibility meters are deployed along the road section at 200-meter intervals, with coordinates at K100+000, K100+200, and K100+400 respectively. A wind field detection device is installed on a gantry at the central location K100+250 of the road section, with the detection coordinates set to (1250, 15, 10), where 1250 is the local horizontal coordinate, 15 is the distance from the road edge, and 10 is the altitude. Vehicle detectors are deployed at the starting point K100+000 and the ending point K100+500 of the road section. The synchronous sampling time is set to 10:00:00:000 milliseconds on November 15, 2023. At this time, the weather sensor at K100+000 measures the temperature. ,humidity ,pressure Visibility measured by the visibility meter The wind field detection device measured the horizontal wind speed. Vertical wind speed Wind angle The average vehicle speed was measured by the vehicle detector at K100+000. Vehicle density Vehicle model number The indication indicates that the vehicles are mainly small cars. The average vehicle speed measured by the vehicle detector at K100+500 is... Vehicle density The vehicle type code indicates the presence of large vehicles. All data is assigned a uniform timestamp. The milliseconds and their corresponding spatial coordinates. The alignment and fusion process associates the environmental data of K100+000 with the traffic flow data at the same location, associates the traffic flow data of K100+500 with its geographical coordinates, and associates the wind field data with the overall spatial coordinates of the road segment. The final generated original monitoring data set entry contains: timestamp 1731639600000, spatial coordinate group ((0,0,0),(1250,15,10),(500,0,0)), environmental data group (8.5,92,1012.3,150), wind field data group (1.2,0.1,45), and traffic flow data group ((22,15,small car owners),(18,25,including large vehicles)).

[0034] S2. Process the original monitoring data set, extract the visibility change trend features and the spatiotemporal distribution features of the wind field, and generate an environmental dynamic feature vector.

[0035] Optionally, the process of processing the original monitoring data set to extract visibility change trend features and wind field spatiotemporal distribution features to generate an environmental dynamic feature vector includes:

[0036] Sliding window difference calculation is performed on the visibility data in the original monitoring data set to extract the visibility decline rate, visibility fluctuation frequency and visibility spatial gradient, and generate visibility change trend features.

[0037] Spatial interpolation and time series decomposition are performed on the three-dimensional wind field data set in the original monitoring data set to extract horizontal wind shear intensity, vertical vorticity and wind direction convergence and divergence to generate the spatiotemporal distribution characteristics of the wind field.

[0038] The visibility change trend features and the wind field spatiotemporal distribution features are normalized and concatenated into vectors to generate an environmental dynamic feature vector that characterizes the dynamic state of environmental change.

[0039] Specifically, the extraction of visibility change trend characteristics is performed from the visibility data sequence in the original monitoring dataset. In the beginning, among them Represents a timestamp. This represents the spatial index of monitoring points. Based on the analysis of 200 sets of measured data on the sudden drop in visibility during the formation of fog patches on highways, the sliding time window length is set. It's 60 seconds long to capture changes at the minute level. At each current moment... For each monitoring point Take time window Visibility data within the area. Rate of visibility decline. Calculate the final visibility value within this window. with initial value The difference, divided by the window duration, yields: Its unit is meters per second, and a negative value indicates a decrease. Visibility fluctuation frequency. By performing a Fast Fourier Transform on the visibility data sequence within the same window, the main frequency components in the power spectrum are extracted. Obtain, that is The dimension is Hertz. Visibility spatial gradient. Then it is necessary to consider adjacent monitoring points. For monitoring points... Find the two nearest monitoring points upstream and downstream. and Calculate the spatial difference of its visibility using the following formula: ,in The mileage coordinates of the monitoring points along the road are given, with the dimension being meters per meter (dimensionless). Their absolute values ​​reflect the spatial non-uniformity of visibility. These three scalar values ​​for each monitoring point are... , , The data are combined to form a sub-vector representing the visibility change trend at that point. The set of sub-vectors from all monitoring points constitutes a complete visibility change trend feature.

[0040] The extraction of spatiotemporal distribution characteristics of wind fields targets three-dimensional wind field datasets. The original data consists of discrete points. Wind speed observation First, determine the wind direction angle. Convert to horizontal wind vector ,in , To obtain the wind field in continuous space, an inverse distance weighting method is used to calculate the horizontal wind vector. and vertical wind speed Spatial interpolation is performed on regular grid points within the target region. For any grid point... Its interpolated wind vector components Calculated as Weight , It is an observation point To grid point The Euclidean distance. and The calculation is similar. Based on the interpolated three-dimensional regular grid wind field... Three features were extracted: horizontal wind shear intensity. At grid points The calculation is based on the horizontal wind speed in the vertical direction. The rate of change on is approximately: ,in Take 2 meters as the unit of measurement, and use units of per second. Vertical vorticity. Describes the intensity of air rotation about a vertical axis in the horizontal plane. The above calculation uses the following formula: Partial derivatives are calculated on the grid using the central difference method, with dimensions in seconds. Wind convergence / divergence. Reflecting the convergence or divergence trend of wind direction, it is calculated as the divergence of the horizontal wind vector. The units are per second. Positive values ​​indicate wind divergence, while negative values ​​indicate convergence. These three scalars are calculated for all key grid points within the target area to form the spatiotemporal distribution characteristics of the wind field.

[0041] The final step in generating dynamic environmental feature vectors is normalization and vector concatenation. Since the dimensions and numerical ranges of each feature differ, normalization is performed based on its physical meaning and the statistical range of historical data, considering both maximum and minimum values. For example, the visibility decay rate... The normalized value is ,in and Based on historical data, it is set as follows and Other characteristic quantities are similarly normalized so that their values ​​fall between 0 and 1. For those with One visibility monitoring point and The system of wind field feature grid points will normalize the visibility change trend feature sub-vector of each monitoring point. The normalized spatiotemporal distribution feature subvector of the wind field at each grid point These elements are concatenated in a fixed spatial order to form a one-dimensional long vector, which is the environmental dynamic feature vector used to characterize the dynamic changes in the environment. Its dimensions are .

[0042] For example, continuing with the target section of the G50 expressway from K100+000 to K100+500, this section has 3 visibility monitoring points, and 5 key grid points are selected for wind field analysis. At the current moment... Sliding window After calculating the visibility data within seconds, the visibility decrease rate was obtained at monitoring point K100+000. Fluctuation frequency The spatial gradient is calculated from the visibility difference between the upstream and downstream points K100+200. (Dimensionless). Similarly, the values ​​for the other two points are obtained. After interpolation of the wind field data, the values ​​at a certain grid point... Calculated horizontal wind shear intensity Vertical vorticity Wind convergence Similarly, the values ​​for the other four grid points are obtained. These values ​​are then normalized according to a preset historical statistical range, for example... Falling Interval, after normalization All normalized scalars are concatenated in a fixed order: first the three monitoring points. There are a total of 9 values, plus 5 grid points. A total of 15 values ​​are collected, ultimately generating a dynamic environmental feature vector containing 24 elements. This vector fully contains the quantitative features extracted from the raw data that characterize the visibility dynamics and wind field structure of the current road segment.

[0043] S3. Based on the environmental dynamic feature vector and the real-time traffic data, construct an environmental energy field-assisted model to calculate the coupling gain coefficient of the expected trajectory of the air cannon, and optimize it based on a preset historical spray event sequence.

[0044] Optionally, the coupling gain coefficient for calculating the expected trajectory of the air cannon based on the environmental dynamic feature vector and the real-time traffic data includes:

[0045] Based on the environmental dynamic feature vector and the real-time traffic flow data set, an environmental energy field assistance model reflecting the interaction between air disturbance propagation and traffic flow motion is constructed.

[0046] Using the environmental energy field-assisted model, the propagation trajectory of air disturbances generated by the air cannon when it sprays at different positions and angles in the environmental energy field is simulated, and the expected action trajectory of the air cannon is generated.

[0047] Based on the matching degree between the expected trajectory of the air cannon and the current dynamic feature vector of the environment, as well as the enhancement or weakening effect of traffic flow on disturbance propagation, a coupling gain coefficient for quantifying the effect of the air cannon is calculated.

[0048] Specifically, the construction of the environmental energy field-assisted model is based on environmental dynamic feature vectors. Based on real-time traffic flow data, the model treats the space surrounding the target road segment as an energy field, and the compressed air mass ejected by the air cannon as an initial disturbance energy packet, whose propagation is influenced by both the background wind field and the moving traffic flow. The core of the model is a description of the disturbance energy density. The evolution equation can be in the form of:

[0049] ,

[0050] in, It is a three-dimensional background wind vector field obtained by interpolating the spatiotemporal distribution features of the wind field from the environmental dynamic feature vector. . It is a traffic flow-induced wind speed field, which is a virtual field calculated from a set of real-time traffic flow data. It is the traffic flow coupling coefficient, used to adjust the intensity of the traffic flow influence. Its dimension is one, and its initial value is calibrated to 0.15 based on computational fluid dynamics simulation and comparison with 200 sets of measured data. This is the turbulent diffusion coefficient, measured in square meters per second. Its value is positively correlated with the horizontal wind shear intensity and vertical vorticity in the spatiotemporal distribution characteristics of the wind field, as determined by empirical formulas. Sure, As the reference diffusion coefficient, take . It is the energy dissipation coefficient, measured per second, and is positively correlated with air humidity, based on humidity data from the basic environmental monitoring dataset. Set as . It is the source term, indicating the position. and time The initial disturbance energy introduced by the air cannon jet has a spatial distribution model based on Gaussian model, and its intensity is proportional to the pressure of the gas storage tank.

[0051] Traffic-induced wind speed field The calculations rely on a real-time traffic flow data set. Vehicles are considered moving obstacles, causing drag and disturbance to the surrounding air. For each vehicle model... (Small car, medium car, large car), assign a characteristic perturbation intensity factor. Their values ​​are 1.0, 1.5, and 2.0 respectively. (At position...) induced wind speed at the location With local vehicle density and average vehicle speed Related, the calculation formula is as follows ,in It is a disturbance factor calculated by weighting the distribution of vehicle models. It is a proportionality constant, with the dimension of meters. The direction of the induced wind is assumed to be the same as the average driving direction of the vehicle.

[0052] Using the aforementioned environmental energy field-assisted model, the process of simulating the expected trajectory of an air cannon involves numerically solving the perturbation energy field. The spatiotemporal evolution. For each air cannon unit. Set its spray position Ejection time , spray direction (pitch angle) Deflection angle and initial energy intensity As input, the model equations are numerically solved using the finite difference method or the finite volume method to calculate the perturbation energy field over a future time period (e.g., 30 seconds). Distribution. Expected trajectory of the air cannon. It is defined as the disturbance energy exceeding a certain threshold. The spatial envelope of the compressed air mass changes over time, visually demonstrating the expected path and extent of its diffusion and impact.

[0053] Calculate the coupling gain coefficient The expected trajectory of action needs to be quantified. The degree of matching with the current environmental conditions and the enhancement effect on traffic flow. Matching degree It is calculated by comparing the difference between the average environmental dynamic characteristics within the trajectory area and the overall average characteristics. ,in It is a trajectory Covering the environmental dynamic feature vectors corresponding to the grid points The average value of the relevant elements in the middle. It is the global average. This is a weight vector used to highlight the impact of key features such as visibility decline rate and wind convergence / divergence. It also reflects the enhancing or weakening effect of traffic flow on disturbance propagation. It is characterized by the ratio of the average traffic-induced wind speed to the background wind speed within the trajectory area, i.e. , This is a small constant added to prevent division by zero. Ultimately, the coupling gain coefficient... Calculated as ,in and It is the weighting coefficient that balances the two contributions, and The initial values ​​are all set to 0.5. It is a dimensionless coefficient. The larger the positive value, the better the expected effect of the air cannon at that position and angle.

[0054] Optionally, the optimization based on a preset historical jetting event sequence includes:

[0055] Identify the periodic fluctuation patterns and abrupt change patterns in the environmental dynamic feature vector to generate environmental evolution pattern identifiers; wherein, the periodic fluctuation patterns consist of regular changes in visibility and wind field parameters, and the abrupt change patterns consist of a sharp decrease in visibility over a short period of time.

[0056] Based on the environmental evolution pattern identifier, historical events with similar pattern identifiers are matched from the preset historical jet event sequence, and the corresponding historical collaborative control instruction set and historical fog elimination effect index are extracted to generate a pattern matching decision reference set.

[0057] Based on the pattern matching decision reference set, the currently calculated coupling gain coefficients are corrected by confidence weighting to generate optimized coupling gain coefficients.

[0058] Specifically, such as Figure 2 As shown, the historical jet event sequence stores the environmental dynamic feature vectors of each past jet event. The collaborative control instruction set adopted, and the fog elimination effect indicators obtained from post-event evaluation. (e.g., visibility improvement rate). The optimization process first identifies the dynamic feature vectors of the current environment. The patterns contained within. Through analysis The system analyzes the changes in visibility trends and the spatiotemporal distribution of wind fields over time to detect the existence of periodic fluctuation patterns or abrupt change patterns. Periodic fluctuation patterns are characterized by regular sinusoidal or cosine variations in parameters such as visibility and wind speed over time; abrupt change patterns are characterized by a sharp, step-like decrease in visibility within a short time window. The system judges the current pattern according to predefined rules and generates an environmental evolution pattern identifier. .

[0059] Next, based on the environmental evolution pattern identification The system retrieves historical events with the same or similar pattern identifiers from the historical jet event sequence. Upon successful matching, it extracts the historical collaborative control command sets corresponding to these historical events and the resulting historical fog elimination effect indicators. This constitutes the pattern matching decision reference set. For the current... For each air cannon unit, identify the event with the best historical performance index from the reference set, and denote the gain coefficient corresponding to the command used at that time as . Its performance indicators are .

[0060] Based on the pattern matching decision reference set, the currently calculated initial coupling gain coefficients Confidence-weighted correction is applied. The optimized coupling gain coefficient is then obtained. for:

[0061] ,

[0062] Among them, weight This is called model confidence, which is based on the characteristics of the current environment. Matching historical environmental characteristics similarity and historical performance indicators The degree of excellence is jointly determined by the similarity. Cosine similarity is used for calculation. The calculation formula is: , This is the theoretical maximum value of the performance indicator. Using this weighting, the optimized coupling gain coefficient... It incorporates both the real-time calculations of the current model and historical successful experiences, thereby generating optimized coupling gain coefficients.

[0063] For example, continuing with the aforementioned target road segment, assume that three air cannon units are deployed. Current environmental dynamic feature vector. Pattern recognition identified it as a "mutation pattern." Five historical events belonging to the same "mutation pattern" were matched from the historical jet event sequence. For air cannon unit 1, the matching degree was calculated after the current model simulated its expected trajectory. Traffic flow effect coefficient Therefore, the initial coupling gain coefficient From historical matching events, find the historical best performance index of Unit 1 when using a certain set of control parameters. (After standardization), the historical gain coefficient at that time Calculate the current environmental characteristics. With regard to the environmental characteristics of this historical event similarity Then the model confidence level Finally, the optimized coupling gain coefficient of unit 1... The calculation process verified all the technical features of the environmental energy field-assisted model construction, the simulation of the expected trajectory, the calculation of the initial coupling gain coefficient, the identification of environmental evolution patterns, the matching of historical events, and the confidence-weighted correction. The data were closely correlated and the technical scope was not expanded.

[0064] S4. Based on the optimized coupling gain coefficient and the visibility change trend characteristics, dynamically generate the collaborative control instruction set for the air gun;

[0065] Optionally, the step of dynamically generating a cooperative control command set for the air cannon based on the optimized coupling gain coefficient and the visibility change trend characteristics includes:

[0066] Based on the visibility decline rate and spatial gradient in the visibility change trend characteristics, determine the core fog areas that require priority intervention and the level of intervention urgency.

[0067] By combining the intervention urgency level with the optimized coupling gain coefficient, the activation sequence, duration of action and spray angle of each air cannon are dynamically allocated to generate an air cannon collaborative control parameter table.

[0068] The air gun cooperative control parameter table is converted into a cooperative control instruction set containing time codes and action parameters that can be recognized by each execution unit in the air gun array.

[0069] Specifically, the core areas of the fog requiring priority intervention and the level of urgency for intervention are determined, based directly on the visibility change trend characteristics generated in step S2, especially the rate of visibility decline. With visibility spatial gradient The core area of ​​the fog is defined as where visibility is below a safe threshold. and For a continuous spatial range of significantly negative values, the safety threshold is... The intervention urgency level is set at 200 meters based on highway driving safety standards. This refers to each monitoring point A comprehensive score considering the speed of fog development and its spatial severity. The calculation formula is as follows:

[0070] ,

[0071] in, This is the maximum setpoint for the rate of visibility decline, used for normalization. ; It represents the absolute value of the spatial gradient of visibility, reflecting the spatial non-uniformity of the fog area; and Let be the weighting coefficient, satisfying Based on statistical analysis of the leading factors of fog diffusion, Set to 0.7. The value is set to 0.3, giving it a higher weight in terms of the rate of change over time. The calculated value is... The values ​​range from 0 to 1 and are divided into three levels of intervention urgency: Due to the high degree of urgency, As medium urgency, The area covered by all high-urgency monitoring points was designated as the core fog zone for priority intervention.

[0072] The activation sequence, duration, and spray angle of each air cannon are dynamically allocated to generate a coordinated control parameter table for the air cannons. This process needs to take into account the level of intervention urgency. And the optimized coupling gain coefficient obtained in step S3 Assuming the air gun array contains Each unit The geographical location is fixed, and its maneuverable spatial range is known. For each marked fog core area, the system retrieves all air cannon units capable of covering that area. Allocation decisions follow the principles of "high urgency priority, high efficiency priority." Activation sequence... The determination is based on the highest urgency level in its target area. , The higher, its The smaller the value, the earlier it starts. Duration of action. Then the coupling gain coefficient of the unit after optimization of the target region And it is positively correlated with the severity of fog in the area, calculated using the following formula: ,in This is the baseline spray duration, set to 3 seconds. The spray angle includes the pitch angle. and horizontal deflection angle These are directly input from the simulation of the expected trajectory of the air cannon in step S3, making... The angle at which the optimal value is achieved is determined. The system iterates through all air cannon units, calculating and inputting parameters for each unit. , , , This ultimately resulted in an air gun collaborative control parameter table, which is a... A data structure with 4 rows and 4 columns.

[0073] Converting the air gun cooperative control parameter table into a cooperative control instruction set is to enable each execution unit in the air gun array to accurately identify and execute actions. The conversion process includes time encoding and parameter encapsulation. Time encoding determines the absolute startup timing. Converted to the time when the system uniform command is issued Delay The unit is milliseconds, i.e. The action parameters include the duration of action. (milliseconds), pitch angle (degrees) and deflection angle (degree). For each air cannon unit This generates a unique, recognizable instruction, which uses a predefined binary or string protocol format. For example, a simple string format instruction would be: "ID:k,Delay:ΔT_k,Duration:T_{dur}(k),Pitch:φ_k,Yaw:φ_k". All A set of instructions, according to Arranging the instructions in ascending order constitutes the final set of coordinated control instructions, ensuring the temporal order of the instructions.

[0074] For example, continuing with the aforementioned target road section and its three air cannon units, the visibility change trend characteristics show that the visibility decrease rate at monitoring point K100+200 is... Spatial gradient Calculate the urgency of intervention. The urgency level is low. At monitoring point K100+400... , ,but The urgency level is medium, and the area near this point is defined as the core region. The optimized coupling gain coefficients are as follows: , , Assume the core area can be covered by air cannon units 2 and 3. Unit 2, due to its... highest and The larger one is assigned to start earliest. Seconds; its duration of action seconds; the injection angle uses the optimal value obtained from simulation, for example , Unit 3 , It is assigned a 1-second delay to start in order to avoid flow field conflicts. seconds; duration of action Seconds. Unit 1 is not assigned any action because it does not cover the core area. This generates the air cannon coordinated control parameter table. During conversion, set... If the system time is 0 milliseconds, then the instruction latency of unit 2 is... milliseconds, unit 3 Milliseconds. The generated cooperative control instruction set contains two instructions, sorted by delay as: "ID:2,Delay:0,Duration:1470,Pitch:10,Yaw:45" and "ID:3,Delay:1000,Duration:1240,Pitch:-5,Yaw:30".

[0075] S5. Perform system security boundary verification and actuator adaptability verification on the cooperative control instruction set, and generate executable drive instructions;

[0076] Optionally, the step of performing system security boundary verification and actuator adaptability verification on the cooperative control instruction set to generate executable drive instructions includes:

[0077] Acquire the system operating status parameters of the air cannon device, including air tank pressure, motor temperature and valve health status, and generate a system safety status vector;

[0078] The collaborative control instruction set is compared with the system security state vector to generate a security verification instruction;

[0079] Based on the mechanical response characteristics and dynamic load capacity of the actuator, the verification safety command is tested for smoothness of action and torque reachability, and an executable drive command adapted to the physical actuator is generated.

[0080] Specifically, acquiring the system operating status parameters of the air cannon device and generating a system safety state vector is the initial step in the verification process. The system operating status parameters are read directly from the embedded sensors and control units of each air cannon unit; core parameters include the air tank pressure. , Winding temperature of the telescopic motor and the health status code of the electrically controlled valve Gas tank pressure The unit is megapascal (MPa), and its safe operating range is set based on pressure vessel design standards. For example, 0.5 MPa to 1.6 MPa. Motor temperature. The unit is Celsius, and its upper limit is... The temperature is set at 120 degrees Celsius based on the motor insulation class. Valve health status. This is an integer code output by the valve controller's self-diagnostic function. 0 represents normal operation, while non-zero values ​​represent different fault types, such as jamming or leakage. For valves... An array of air cannon units, each unit These three parameters are combined into a subvector. All The subvectors are concatenated in unit number order to form a dimension of ... System security state vector This vector reflects the physical health of the entire jet array in real time.

[0081] Comparing the cooperative control instruction set with the system security state vector to generate verification security instructions is a process of security filtering instruction by instruction. The cooperative control instruction set comes from step S4, where each instruction specifies a target unit. Delay time, duration of action, pitch angle, and yaw angle. During comparison, the system retrieves the target unit from the command. In the system security state vector The corresponding parameter segment. The verification rules are based on preset safety boundary conditions, expressed as inequalities. For example, for the pressure of the gas storage tank, the requirements are... If the requirements are not met, the unit is deemed unusable for injection. Regarding motor temperature, the requirements are... If the value exceeds a certain limit, the motor will be prohibited from performing angle adjustment actions. Regarding the valve's health status, the requirements are as follows: A non-zero value indicates a valve malfunction and prohibits its opening. Only when the target unit... The original instruction is marked as "safe" and retained as is only if all state parameters satisfy their corresponding safety boundary conditions. If any condition is not met, the system generates an alternative "safety verification instruction," which forcibly modifies the duration to zero and sets the pitch and yaw angles to safe default positions. This cancels the unit's current action while maintaining the integrity of the instruction structure to ensure timing synchronization. After traversing the entire cooperative control instruction set, the output is a safety verification instruction set, filtered for safety, consisting of the original instruction marked "safe" and the alternative "cancel action" instruction.

[0082] Based on the mechanical response characteristics and dynamic load capacity of the actuator, the smoothness of action and torque availability of the safety command are verified, which is a further refinement of the feasibility of the "safety" command. As actuators, telescopic motors and electrically controlled valves have physical limits to their movement. The smoothness of action verification aims to prevent the motor from undergoing abrupt, violent movements when adjusting the angle, thus avoiding mechanical shock. The verification method is to calculate two adjacent control cycles (let the cycle be 1). Within ) the amount of angle change required by the instruction. and The maximum permissible angular velocity of the motor Based on the motor specifications, for example, 30 degrees per second. Smoothness requirements. and For commands that do not meet the conditions, the system will deselect their angle parameters. and Corrected to be from the position of the previous time step Speed ​​at The position that can be reached within the range is the rate limit.

[0083] Torque availability verification ensures that the motor has sufficient torque to drive the load (nozzle and its associated mechanisms) to the required angle. Required torque. Moment of inertia relative to the load angular acceleration and frictional torque Related, can be estimated as Angular acceleration Derived from the angle change curve required by the command. The continuous output torque of the motor at a specific speed. It is a known characteristic curve. The verification condition is: If this condition is not met, it indicates that the target angle cannot be accurately reached under the current dynamic conditions. In this case, the system will use an iterative backtracking method to gradually reduce the angle change range required by the command. and Until the estimated satisfy The constraints are then used to generate a new set of torque-achievable angle parameters. After verification and correction for smoothness and torque availability, the safety instructions are transformed into a set of executable drive instructions fully adapted to the motion capabilities and safety limitations of the physical actuator. These instructions retain the original time coding, but the motion parameters have been fine-tuned to ensure safety, smoothness, and power availability.

[0084] For example, suppose the air gun array contains 3 units, and the cooperative control instruction set from step S4 contains 3 instructions. System safety state vector Real-time reading: Pressure of Unit 1 gas storage tank MPa (within the safe range of 0.5-1.6 MPa), motor temperature Valve health code (temperature below 120 degrees Celsius) (Normal). Unit 2 pressure is 1.7 MPa (exceeding normal). The temperature is 95 degrees Celsius, and the valve code is 0. Unit 3 has a pressure of 1.4 MPa and a temperature of 121 degrees Celsius (exceeding...). The valve code is 2 (fault). After comparison, the status of Unit 1 is all qualified, and its instructions are marked as "safe" and retained. Unit 2 is due to excessive pressure, and Unit 3 is due to excessive temperature and valve fault. Their original instructions are replaced with "cancel action" instructions, with an action duration of 0 and the angle set to the default value. This generates a set of verification safety instructions.

[0085] Next, the smoothness of the "safety" command in Unit 1 was verified. This command requires the pitch angle to be adjusted from the current 5 degrees to 20 degrees, with a control cycle of... If the time is seconds, then the required angular velocity is... degrees / second, exceeding The system has a limit of degrees per second. Therefore, the system corrects the target pitch angle to... Degree. Torque achievability verification estimates the torque required for the current motion. The torque is 12 N·m. Consulting the motor characteristic curve, we know that it can continuously output torque at the current speed. 10 N·m Therefore, the condition is not met. The system iterates back, halving the angle change, and recalculates. It is 9 Newton-meters, at which point it satisfies Therefore, the target pitch angle was further corrected to Ultimately, the executable drive commands for Unit 1 were determined as follows: the delay time and duration remained unchanged, the pitch angle was corrected to 6.5 degrees, and the yaw angle underwent a similar correction. The output commands for Units 2 and 3 were cancellation commands.

[0086] S6. Based on the executable drive command, drive the telescopic motor and the electronically controlled valve to perform coordinated actions to trigger the air cannon to spray;

[0087] Optionally, the step of driving the telescopic motor and the electronically controlled valve to perform coordinated actions based on the executable drive command to trigger the air cannon spray includes:

[0088] The executable drive command is parsed, and the stroke control parameters for the telescopic motor and the opening timing parameters for the electrically controlled valve are extracted to generate motor control signals and valve control signals.

[0089] Based on the motor control signal, drive the telescopic motor to adjust the pitch and deflection angles of the air cannon nozzle to the target position;

[0090] Synchronously, based on the valve control signal, the electronically controlled valve is opened according to the specified timing and opening degree, releasing compressed air and triggering the air cannon to spray along the target angle and trajectory.

[0091] Specifically, parsing the executable drive instructions and extracting the stroke control parameters for the telescopic motor and the opening timing parameters for the electrically controlled valve is the starting point for action execution. The set of executable drive instructions comes from step S5, where each instruction includes a target unit identifier and a delay time. Duration of action Pitch angle and deflection angle The analysis process is completed by the local controller of each air cannon unit. The stroke control parameters mainly refer to the target angular position that the motor needs to achieve. and And optional motion speed curves, such as using an S-shaped acceleration / deceleration curve to avoid impact. The opening timing parameter is a function of time, defining the valve's... After the time delay, at what speed and to what extent should the opening be initiated and maintained? Close after a set time. The opening speed is typically determined by the valve's response time constant. The decision, and the maximum opening. The required airflow is then set, for example, 100% represents fully open. After extracting these key parameters from the command, the controller converts them into motor control signals and valve control signals respectively. The motor control signal is typically a position command pulse sequence or analog voltage signal sent to the servo driver or stepper motor driver. The valve control signal is typically a pulse width modulation signal or current signal sent to the proportional valve or high-speed switching valve.

[0092] Based on the motor control signal, the telescopic motor is driven to adjust the pitch and yaw angles of the air cannon nozzle to the target position, involving closed-loop position control. The telescopic motor typically refers to a linear motor or electric actuator, whose linear displacement... The pitch angle of the nozzle is converted through a linkage mechanism. The relationship is ,function Determined by the specific mechanical design, for example ,in This refers to the length of the rotating arm. The controller will set the target angle. and The target stroke required for each motor is calculated using an inverse kinematics model. and Subsequently, the controller compares the current stroke fed back by the motor encoder. With the target itinerary Generate error signal The error signal is input to a proportional-integral-derivative controller, and its output formula is:

[0093] ,

[0094] in These are control quantities, such as voltage or pulse width, output to the motor. , , These are the proportional, integral, and derivative gain coefficients, which are determined experimentally based on the motor's response characteristics and load inertia. The controller continuously adjusts the output until the error... Once the nozzle enters the preset dead zone range, such as ±0.5 mm, the nozzle angle is considered to be properly adjusted. To ensure repeatability, the system performs zero-point calibration and stroke limit calibration every time it is powered on or periodically.

[0095] Synchronizing the opening of electrically controlled valves according to a specified timing and degree of opening based on valve control signals is crucial for achieving compressed air release. Electrically controlled valves typically employ high-speed solenoid valves or proportional control valves. The opening timing parameters are converted into specific valve drive signals. For the simplest on / off valve, It is in time The voltage jumps from 0 to the rated voltage (e.g., 24V) and... The square wave signal transitions back to 0 at time jumps. For a proportional valve used to achieve flow regulation, then... It could be a target opening degree A proportional analog current signal (e.g., 4-20mA). During the startup phase, the temperature rises from 0 to [a certain value] according to a preset curve. After receiving the signal, the valve controller actuates the valve core, thereby changing the flow area and releasing compressed air from the air tank. The valve opening response time... The response time for closing must be much smaller than This ensures that the shape of the air jet matches the expected shape. Crucially, the valve opening trigger must be precisely synchronized with the moment the telescopic motor completes its angle adjustment. This is typically achieved through a hardware timer interrupt on the controller, ensuring that the valve action is triggered within a very short time (e.g., within 10 milliseconds) after the motor reaches its position signal, thereby ultimately triggering the air cannon to spray along the target angle and trajectory.

[0096] For example, taking the aforementioned modified unit 1 executable driver instruction as an example, its content is: ID:1, delay milliseconds, duration milliseconds, pitch angle Degrees, deflection angle The local controller parses the instruction. For the telescopic motor, the target travel is calculated using an inverse kinematics model. Assuming the current pitch angle is 5 degrees, the corresponding travel... Millimeters, target 6.5 degrees corresponding The controller uses a PID algorithm to generate a pulse sequence to drive the pitch motor until the encoder feedback travel reaches a range of 65±0.5 mm, taking approximately 200 milliseconds. The yaw motor performs a similar adjustment. The position sensor sends a signal indicating that the angle adjustment is complete. For electrically controlled valves, the controller starts a hardware timer at the beginning of command parsing, set at... The valve is triggered immediately upon receiving a motor position signal within milliseconds (i.e., instantly). It generates a square wave signal with a duration of 1470 milliseconds and an amplitude of 24V. The high-speed solenoid valve is activated. Upon receiving the signal, the valve fully opens within approximately 15 milliseconds, and compressed air is ejected from the nozzle along a pre-adjusted 6.5-degree pitch angle and 30-degree yaw angle. After 1470 milliseconds, the controller will... Set to zero, close the valve, and stop spraying.

[0097] S7. After the air cannon is sprayed, the original monitoring data set is updated to evaluate the fog elimination effect and fed back to the environmental energy field assist model for correction.

[0098] Optionally, the step of updating the original monitoring data set to evaluate the fog removal effect after the air cannon is fired, and feeding it back to the environmental energy field-assisted model for correction, includes:

[0099] After the air cannon spraying action is completed, multi-source environmental data and real-time traffic flow data of the target road section are reacquired, and the original monitoring data set is updated.

[0100] By comparing the updated set of original monitoring data with the set of original monitoring data before spraying, the visibility improvement rate, fog dissipation area and wind field structure change are calculated to generate fog elimination effect indicators.

[0101] The fog elimination effect index, the corresponding expected trajectory of the air cannon, and the environmental dynamic feature vector are correlated to form training samples and update the historical spray event sequence.

[0102] The parameters in the environmental energy field-assisted model are iteratively optimized using the updated historical jet event sequence to complete dynamic feedback correction.

[0103] Specifically, reacquiring multi-source environmental data and real-time traffic flow data for the target road segment and updating the original monitoring data set is the foundation for effect evaluation. After all the coordinated air cannon spraying actions triggered in step S6 are completed, the system waits for a preset effect observation period. The physical timescale setting based on the diffusion of compressed air masses and the dissipation of fog particles is typically 60 to 120 seconds. After the observation period ends, the system immediately repeats the same multi-source data acquisition and fusion process as step S1. At this time, meteorological sensors, wind field detection devices, and vehicle detectors... A new round of data is collected. This data is aligned with the same timestamps and fused with spatial coordinates to generate a raw monitoring dataset with spatiotemporal labels after the spraying process. This set is structurally similar to the original monitoring data set prior to injection. They are completely identical, thus enabling direct comparison.

[0104] By comparing the updated set of original monitoring data with the set of original monitoring data before spraying, the visibility improvement rate, fog dissipation area, and wind field structure change were calculated, which constituted the indicators of fog dissipation effectiveness. The quantification process. Visibility improvement rate. Focusing on the average improvement in visibility at specific monitoring points or areas. For each visibility monitoring point... Calculate its visibility change value ,in and These represent the visibility values ​​at that point before and after spraying, respectively. The overall visibility improvement rate is also shown. Then it is defined as all monitoring points The average value and a reference visibility The ratio of (e.g., initial average visibility or safety threshold) is expressed by the formula: Its dimension is one, representing a percentage increase.

[0105] Fog dissipation area The assessment evaluates the mitigation of fog within a spatial range. The target road segment is discretized into a regular grid on a horizontal plane. Based on... and Spatial interpolation results of medium visibility data, determining the value of each grid point. Conditions before and after spraying: Visibility value (If the safety threshold is 200 meters), then mark that point as a "fog-affected point". Fog dissipation area That is to be in The middle is marked as "foggy" and in The total area of ​​all grid points transformed into "fog-free points" is calculated using the following formula: .in, It is an indicator function, when the point The value is 1 when the above state transition conditions are met, and 0 otherwise. It is the area of ​​each grid cell, in square meters.

[0106] Changes in wind field structure The aim is to quantify the degree of disturbance to the local wind field caused by air cannon spray. It calculates this by comparing the differences between three-dimensional wind field data sets before and after spraying. First, the wind field data before and after spraying are interpolated onto the same spatial grid to obtain the horizontal wind vector at each grid point. and Changes in wind field structure Defined as the root mean square value of the magnitude of the wind vector change at these grid points, its calculation formula is:

[0107] ,

[0108] in, This represents the total number of grid points involved in the calculation. The unit is meters per second; the larger the value, the more significant the change in the background wind field caused by the air cannon. The calculated value... , , The three scalars are normalized to fall within the range of 0 to 1, and then combined into a three-dimensional vector to generate a complete index of fog removal effectiveness. .

[0109] Associating fog dissipation effectiveness indicators, the corresponding expected trajectory of the air cannon, and environmental dynamic feature vectors to form training samples and update historical spray event sequences is key to accumulating experience. A training sample is a data packet containing the following elements: the environmental dynamic feature vector before spraying, which serves as input. The set of expected action trajectories of the air cannon generated in step S3 serves as the basis for system decision-making and prediction. and the corresponding optimized set of coupling gain coefficients The newly calculated index of fog removal effect is the actual output. This data packet also includes a copy of the actual executed set of collaborative control instructions. Each packet is assigned a unique event ID and timestamp. The system adds this as a new entry to a pre-defined historical jet event sequence database. This sequence is a time-sorted structured dataset that grows with each jet event, forming an experience base for model optimization.

[0110] Iterative optimization of the parameters in the environmental energy field-assisted model using the updated historical injection event sequence is the final step in completing the dynamic feedback correction. The environmental energy field-assisted model includes several adjustable parameters, such as the vehicle-flow coupling coefficient. The weighting coefficients in the turbulent diffusion coefficient formula and the weights in the coupling gain coefficient calculation. The goal of the optimization process is to maximize the correlation between the model's predicted gain coefficients and the final actual results. A loss function is defined. For example, the mean squared error function is used to measure the performance of the first historical sequence. The overall effect predicted by the model for each event (by...) Comprehensive characterization and actual effect The differences between them are discussed. Using the entire updated historical sequence data, gradient descent or its variants are employed to iteratively adjust the model parameters to minimize the total loss. With each parameter update, the model's simulation of the expected trajectory of the future air cannon and the calculation of the coupling gain coefficient will become more accurate, thus completing the backpropagation and correction from the actual effect to the model parameters, forming a self-improving closed loop.

[0111] For example, continuing with the aforementioned target road segment example. After spraying is performed in units 1 and 3, a waiting observation period is observed. Seconds. Then, the system re-collects data and generates... Calculations showed that visibility at monitoring point K100+400 improved from 80 meters to 180 meters. Meters. Average visibility improved at all monitoring points. meters, reference visibility Meters, then visibility improvement rate (i.e., 30%). Spatial grid analysis shows that an area of ​​3500 square meters changed from a foggy state to a fog-free state, therefore the fog dissipation area... After normalization, it becomes 0.35. Wind field change. Calculated as After normalization, it becomes 0.25. Therefore, the index for fog removal effectiveness is... Please use this indicator. The dynamic feature vector of the environment prior to this event Simulated expected trajectory of the air cannon This data, along with the actual executable drive commands issued, is packaged into a training sample and stored in a historical injection event sequence. Assuming there are already 99 events in the sequence, this sample becomes the 100th. Subsequently, the optimization algorithm uses these 100 samples to adjust the parameters in the environmental energy field-assisted model, such as adjusting the vehicle flow coupling coefficient. The value was fine-tuned from 0.15 to 0.16, so that the model's predicted gain coefficient was more consistent with historical actual performance metrics. The correlation increased.

[0112] Based on the same inventive concept, such as Figure 3 As shown, the present invention also provides a highway fog elimination system based on the principle of an air cannon, the system comprising:

[0113] The environmental traffic flow perception module is used to acquire multi-source environmental data and real-time traffic flow data of the target road section and generate a raw monitoring data set;

[0114] The feature extraction module is used to process the original monitoring data set, extract visibility change trend features and wind field spatiotemporal distribution features, and generate environmental dynamic feature vectors.

[0115] The coupling analysis module is used to construct an environmental energy field-assisted model based on the environmental dynamic feature vector and the real-time traffic data to calculate the coupling gain coefficient of the expected trajectory of the air cannon, and to optimize it based on a preset historical spray event sequence.

[0116] The intelligent decision-making module is used to dynamically generate a set of coordinated control instructions for the air cannon based on the optimized coupling gain coefficient and the visibility change trend characteristics.

[0117] The instruction generation module is used to perform system security boundary verification and actuator adaptability verification on the cooperative control instruction set, and generate executable drive instructions.

[0118] The instruction execution module is used to drive the telescopic motor and the electronically controlled valve to perform coordinated actions based on the executable drive instructions, thereby triggering the air cannon to spray.

[0119] The feedback correction module is used to update the original monitoring data set after the air cannon is sprayed to evaluate the fog elimination effect and feed it back to the environmental energy field assist model for correction.

[0120] It should be noted that the above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the invention. All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of the invention upon considering the disclosure of the specification and practical truths. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A method for eliminating highway fog based on the principle of an air cannon, characterized in that, The method includes: Acquire multi-source environmental data and real-time traffic flow data of the target road segment to generate a raw monitoring data set; The original monitoring data set is processed to extract visibility change trend features and wind field spatiotemporal distribution features, and an environmental dynamic feature vector is generated. Based on the environmental dynamic feature vector and the real-time traffic flow data, an environmental energy field-assisted model is constructed to calculate the coupling gain coefficient of the air cannon's expected trajectory, and optimized based on a preset historical spray event sequence. The calculation of the coupling gain coefficient includes: constructing an environmental energy field-assisted model reflecting the interaction between air disturbance propagation and traffic flow movement based on the environmental dynamic feature vector and the real-time traffic flow data set; using the environmental energy field-assisted model, simulating the propagation trajectory of air disturbance generated by the air cannon at different positions and angles in the environmental energy field to generate the expected trajectory of the air cannon; and calculating the coupling gain coefficient used to quantify the air cannon's effect based on the matching degree between the expected trajectory of the air cannon and the current environmental dynamic feature vector, as well as the enhancement or weakening effect of traffic flow movement on disturbance propagation. Based on the optimized coupling gain coefficient and the visibility change trend characteristics, a collaborative control command set for the air cannon is dynamically generated. The system security boundary verification and actuator adaptability verification are performed on the cooperative control instruction set to generate executable drive instructions; Based on the executable drive command, the telescopic motor and the electronically controlled valve are driven to perform coordinated actions to trigger the air cannon to spray. After the air cannon is fired, the original monitoring data set is updated to evaluate the fog removal effect and fed back to the environmental energy field-assisted model for correction.

2. The method for eliminating highway fog based on the principle of an air cannon according to claim 1, characterized in that, The process of acquiring multi-source environmental data and real-time traffic flow data of the target road segment to generate the original monitoring data set includes: Acquire temperature and humidity data and atmospheric pressure data collected by meteorological sensors installed along the target road section, and real-time visibility data collected by visibility meters, to generate a basic environmental monitoring data set; Acquire horizontal wind speed, vertical wind speed, and wind direction data collected by a wind field detection device set above the target road section, and generate a three-dimensional wind field data set; The system acquires vehicle speed, vehicle density, and vehicle type data collected by vehicle detectors within the target road segment, and generates a real-time traffic flow data set. The basic environmental monitoring data set, the three-dimensional wind field data set, and the real-time traffic flow data set are aligned and fused according to a unified timestamp and spatial coordinates to generate an original monitoring data set with spatiotemporal labels.

3. The method for eliminating highway fog based on the principle of an air cannon according to claim 2, characterized in that, The process of processing the original monitoring data set to extract visibility change trend features and wind field spatiotemporal distribution features, and generating an environmental dynamic feature vector includes: Sliding window difference calculation is performed on the visibility data in the original monitoring data set to extract the visibility decline rate, visibility fluctuation frequency and visibility spatial gradient, and generate visibility change trend features. Spatial interpolation and time series decomposition are performed on the three-dimensional wind field data set in the original monitoring data set to extract horizontal wind shear intensity, vertical vorticity and wind direction convergence and divergence to generate the spatiotemporal distribution characteristics of the wind field. The visibility change trend features and the wind field spatiotemporal distribution features are normalized and concatenated into vectors to generate an environmental dynamic feature vector that characterizes the dynamic state of environmental change.

4. The method for eliminating highway fog based on the principle of an air cannon according to claim 3, characterized in that, The optimization based on the preset historical jet event sequence includes: Identify the periodic fluctuation patterns and abrupt change patterns in the environmental dynamic feature vector to generate environmental evolution pattern identifiers; wherein, the periodic fluctuation patterns consist of regular changes in visibility and wind field parameters, and the abrupt change patterns consist of a sharp decrease in visibility over a short period of time. Based on the environmental evolution pattern identifier, historical events with similar pattern identifiers are matched from the preset historical jet event sequence, and the corresponding historical collaborative control instruction set and historical fog elimination effect index are extracted to generate a pattern matching decision reference set. Based on the pattern matching decision reference set, the currently calculated coupling gain coefficients are corrected by confidence weighting to generate optimized coupling gain coefficients.

5. The method for eliminating highway fog based on the principle of an air cannon according to claim 4, characterized in that, The step of dynamically generating a coordinated control instruction set for the air cannon based on the optimized coupling gain coefficient and the visibility change trend characteristics includes: Based on the visibility decline rate and spatial gradient in the visibility change trend characteristics, determine the core fog areas that require priority intervention and the level of intervention urgency. By combining the intervention urgency level with the optimized coupling gain coefficient, the activation sequence, duration of action and spray angle of each air cannon are dynamically allocated to generate an air cannon collaborative control parameter table. The air gun cooperative control parameter table is converted into a cooperative control instruction set containing time codes and action parameters that can be recognized by each execution unit in the air gun array.

6. The method for eliminating highway fog based on the principle of an air cannon according to claim 5, characterized in that, The step of performing system security boundary verification and actuator adaptability verification on the collaborative control instruction set to generate executable driver instructions includes: Acquire the system operating status parameters of the air cannon device, including air tank pressure, motor temperature and valve health status, and generate a system safety status vector; The collaborative control instruction set is compared with the system security state vector to generate a security verification instruction; Based on the mechanical response characteristics and dynamic load capacity of the actuator, the verification safety command is tested for smoothness of action and torque reachability, and an executable drive command adapted to the physical actuator is generated.

7. The method for eliminating highway fog based on the principle of an air cannon according to claim 6, characterized in that, The step of driving the telescopic motor and the electronically controlled valve to perform coordinated actions based on the executable drive command to trigger the air cannon spray includes: The executable drive command is parsed, and the stroke control parameters for the telescopic motor and the opening timing parameters for the electrically controlled valve are extracted to generate motor control signals and valve control signals. Based on the motor control signal, drive the telescopic motor to adjust the pitch and deflection angles of the air cannon nozzle to the target position; Synchronously, based on the valve control signal, the electronically controlled valve is opened according to the specified timing and opening degree, releasing compressed air and triggering the air cannon to spray along the target angle and trajectory.

8. The method for eliminating highway fog based on the principle of an air cannon according to claim 7, characterized in that, The process of updating the original monitoring data set after air cannon spraying to evaluate the fog removal effect and feeding it back to the environmental energy field-assisted model for correction includes: After the air cannon spraying action is completed, multi-source environmental data and real-time traffic flow data of the target road section are reacquired, and the original monitoring data set is updated. By comparing the updated set of original monitoring data with the set of original monitoring data before spraying, the visibility improvement rate, fog dissipation area and wind field structure change are calculated to generate fog elimination effect indicators. The fog elimination effect index, the corresponding expected trajectory of the air cannon, and the environmental dynamic feature vector are correlated to form training samples and update the historical spray event sequence. The parameters in the environmental energy field-assisted model are iteratively optimized using the updated historical jet event sequence to complete dynamic feedback correction.

9. A highway fog elimination system based on the principle of an air cannon, characterized in that: The system includes: The environmental traffic flow perception module is used to acquire multi-source environmental data and real-time traffic flow data of the target road section and generate a raw monitoring data set; The feature extraction module is used to process the original monitoring data set, extract visibility change trend features and wind field spatiotemporal distribution features, and generate environmental dynamic feature vectors. The coupling analysis module is used to construct an environmental energy field-assisted model based on the environmental dynamic feature vector and the real-time traffic flow data to calculate the coupling gain coefficient of the expected trajectory of the air cannon, and to optimize it based on a preset historical spray event sequence. The calculation of the coupling gain coefficient includes: constructing an environmental energy field-assisted model reflecting the interaction between air disturbance propagation and traffic flow movement based on the environmental dynamic feature vector and the real-time traffic flow data set; using the environmental energy field-assisted model, simulating the propagation trajectory of air disturbance generated by the air cannon at different positions and angles in the environmental energy field to generate the expected trajectory of the air cannon; and calculating the coupling gain coefficient used to quantify the effect of the air cannon based on the matching degree between the expected trajectory of the air cannon and the current environmental dynamic feature vector, as well as the enhancement or weakening effect of traffic flow on disturbance propagation. The intelligent decision-making module is used to dynamically generate a set of coordinated control instructions for the air cannon based on the optimized coupling gain coefficient and the visibility change trend characteristics. The instruction generation module is used to perform system security boundary verification and actuator adaptability verification on the cooperative control instruction set, and generate executable drive instructions. The instruction execution module is used to drive the telescopic motor and the electronically controlled valve to perform coordinated actions based on the executable drive instructions, thereby triggering the air cannon to spray. The feedback correction module is used to update the original monitoring data set after the air cannon is sprayed to evaluate the fog elimination effect and feed it back to the environmental energy field assist model for correction.

Images