UAV cruise warning method and monitoring system for high-speed traffic monitoring

Abstract

The present invention provides a drone cruise pre-warning method and monitoring system for high-speed traffic monitoring. The method collects dynamic monitoring data of a target road section in real time through multiple hovering monitoring platforms deployed along the highway. Based on the dynamic monitoring data, it identifies abnormal events in the target road section and generates multi-level warning signals according to the event type of the abnormal event. These signals are transmitted to a command center, which generates drone dispatch instructions that match the event type. The target drone is controlled to move to the target location corresponding to the abnormal event based on path planning parameters. The target drone's external module executes warning operations corresponding to the event type. The monitoring range of the hovering monitoring platform is adjusted according to the duration of the abnormal event, and traffic flow regulation instructions are sent to the service area control system of vehicles in the target direction. The present invention can improve the timeliness of accident warnings and the systematic nature of disposal plans, effectively reducing the incidence of secondary accidents and optimizing the efficiency of emergency resource utilization.

Description

Technical Field

[0001] The present invention relates to the fields of road monitoring and traffic control, and in particular to a drone cruise warning method and a monitoring system for high-speed traffic monitoring. Background Art

[0002] In the field of intelligent highway monitoring, research on traffic accident warning and emergency response technologies continues to deepen. Existing technologies primarily rely on a combination of ground-based sensor networks and manual inspections. Fixed cameras collect road condition data, which is then analyzed and processed by a command center before being disseminated via navigation systems. However, this approach faces challenges such as limited monitoring coverage due to sensor density, delayed manual response times that increase the risk of secondary accidents, and the one-way transmission of warning information, which lacks the ability to proactively intervene on-site. Especially over long stretches, achieving instantaneous detection and three-dimensional response to accident locations is difficult, leading to inefficient emergency lane management and delayed traffic flow control in service areas. Traditional drone applications are often limited to single-point aerial photography and evidence collection, lacking deep integration with monitoring systems for energy supply, signal relay, and path planning. This results in low utilization of aerial warning resources and an inability to establish a comprehensive response chain covering the entire accident lifecycle, hindering the intelligent upgrade of highway safety and control systems. Summary of the Invention

[0003] The present invention provides a UAV cruise warning method and system for high-speed traffic monitoring.

[0004] In a first aspect, an embodiment of the present invention provides a drone cruise warning method for high-speed traffic monitoring, comprising:

[0005] The system collects dynamic monitoring data of the target road section in real time through multiple hovering monitoring platforms deployed along the highway, wherein each of the hovering monitoring platforms includes an energy supply module and a signal relay module;

[0006] Identifying abnormal events existing in a target road section based on the dynamic monitoring data, and generating multi-level warning signals according to the event type of the abnormal event;

[0007] Transmitting the multi-level warning signal to a command center, triggering the command center to generate a drone dispatch instruction matching the event type, the drone dispatch instruction including the target drone's identifier and path planning parameters;

[0008] Controlling the target UAV to move to a target location corresponding to the abnormal event based on the path planning parameters, and executing a warning operation corresponding to the event type through the external speaker module of the target UAV;

[0009] The monitoring range of the hovering monitoring platform is adjusted according to the persistence of the abnormal event, and a flow regulation instruction is sent to the service area control system of the target direction vehicle to dynamically control the traffic strategy of the service area entrance.

[0010] In a second aspect, an embodiment of the present invention provides a monitoring system, including:

[0011] a memory storing a computer program;

[0012] A processor is used to load the computer program to implement the drone cruise warning method for high-speed traffic monitoring as described above.

[0013] The drone cruise warning method for high-speed traffic monitoring provided by the present invention realizes wide-area dynamic monitoring of target road sections by deploying a hovering monitoring platform with integrated energy supply and signal relay functions, generates multi-level warning signals that accurately match the event type based on multi-dimensional feature analysis, and realizes rapid and accurate response through path parameterized control of drone scheduling instructions, forming an energy-signal-space collaborative system between the high-altitude monitoring platform and the drone; while the drone performs directional warning operations, it dynamically reconstructs the coverage of the monitoring network according to the event diffusion situation, and links with the service area control system to implement traffic flow prediction and diversion, breaking through the limitations of traditional manual warnings such as large monitoring blind spots, high response delays, and fragmented secondary disaster prevention and control. Through multi-level linkage of event perception, resource scheduling, on-site disposal and road network regulation, without increasing hardware complexity, the timeliness of accident warnings and the systematic nature of disposal plans are significantly improved, effectively reducing the incidence of secondary accidents and optimizing the efficiency of emergency resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 This is a flow chart of a drone cruise warning method for high-speed traffic monitoring provided by an embodiment of the present invention.

[0016] Figure 2 It is a schematic diagram of the composition of a monitoring system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 , Figure 1 A flowchart of a drone cruise warning method for high-speed traffic monitoring provided in an embodiment of the present invention is provided. The drone cruise warning method for high-speed traffic monitoring can be executed by a monitoring system and may include the following steps:

[0019] Step S100: Dynamic monitoring data of a target road section is collected in real time by deploying multiple hovering monitoring platforms along the highway, wherein each hovering monitoring platform includes an energy supply module and a signal relay module.

[0020] A hovering monitoring platform can be implemented using a towed hot air balloon. Along highways, a towed hot air balloon is installed at predetermined intervals (e.g., 100 kilometers) as a hovering monitoring platform. A drone slot can be installed on top of the balloon for docking drones. Furthermore, the balloon's connection to the ground is via a conduit, allowing it to function as a UPS power supply for charging drones. The energy supply module provides power to the entire hovering monitoring platform and the docked drones, ensuring normal operation. The signal relay module relays and transmits signals, enhancing signal coverage and strength.

[0021] In one embodiment, a wide-area monitoring module is configured within a preset altitude range of a hovering monitoring platform. The wide-area monitoring module includes a multi-directional sensor array and a satellite communication unit. Based on this, step S100, where multiple hovering monitoring platforms are deployed along the highway to collect dynamic monitoring data of the target road section in real time, may specifically include the following steps S110-S140:

[0022] Step S110: Perform omnidirectional scanning on the target road section through a multi-directional sensor array to obtain real-time image data and environmental parameter data of the target road section.

[0023] A multi-directional sensor array consists of multiple sensors in different directions, mounted within a preset height range on a hovering monitoring platform. These sensors can include cameras and meteorological sensors. The cameras are used to obtain real-time image data of the target road section, while the meteorological sensors are used to collect environmental parameter data such as temperature, humidity, and wind speed. In practical applications, the multi-directional sensor array performs omnidirectional scans of the target road section at regular intervals, ensuring comprehensive and real-time acquisition of relevant information about the target road section. For example, on a clear weekday morning, the multi-directional sensor array scans a section of highway. The cameras continuously capture vehicle traffic on the section, while the meteorological sensors simultaneously record ambient temperature, wind speed, and other parameters of the section, thereby obtaining real-time image data and environmental parameter data for the target road section.

[0024] Step S120: inputting the real-time image data into a preset event recognition model, outputting the dynamic trajectory characteristics of the vehicle behavior in the target road section, and comparing the dynamic trajectory characteristics with a preset safety threshold.

[0025] The pre-set event recognition model can utilize a deep learning model, such as a convolutional neural network (CNN). This model, trained with a large amount of annotated data, can accurately identify vehicle behavior characteristics in image data. Convolutional neural networks typically consist of an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. The input layer receives real-time image data, the convolutional layer extracts image features using convolution kernels, the pooling layer performs dimensionality reduction on these features, the fully connected layer integrates the extracted features, and the output layer outputs the dynamic trajectory characteristics of the vehicle's behavior.

[0026] To train this model, a large amount of highway image data is first collected and vehicle behavior is annotated, such as speed, direction, and lane changes. This annotated data is then divided into a training set and a validation set. The convolutional neural network is trained using the training set, and the model parameters are continuously adjusted using a backpropagation algorithm to minimize the error between the model's output and the annotated results. The trained model is then validated on the validation set to evaluate its performance.

[0027] In practical applications, the real-time image data acquired in step S110 is input into a trained event recognition model, which then outputs dynamic trajectory features of vehicle behavior on the target road section, such as vehicle trajectory and speed changes. These dynamic trajectory features are then compared with preset safety thresholds, which can be set based on extensive historical data and traffic regulations, such as the normal speed range and reasonable frequency of lane changes.

[0028] Step S130: If the dynamic trajectory feature exceeds the safety threshold, a preliminary warning signal is sent to the command center via the satellite communication unit, and an abnormal timestamp is marked in the local storage module of the hovering monitoring platform.

[0029] The satellite communication unit is responsible for transmitting the data collected and the signals generated by the hovering monitoring platform via satellite to ensure that the signals can be sent to the command center in a timely and accurate manner. When the comparison result in step S120 shows that the dynamic trajectory characteristics exceed the safety threshold, it indicates that there may be abnormal conditions in the target road section, such as vehicle speeding, frequent lane changes, etc. At this time, the hovering monitoring platform sends a preliminary alarm signal to the command center through the satellite communication unit to inform the command center that there may be abnormalities in the road section. At the same time, the abnormal timestamp is marked in the local storage module of the hovering monitoring platform, and the time when the abnormal event occurred is recorded to facilitate subsequent data analysis and processing. For example, when the driving speed of a vehicle exceeds the preset safety speed threshold, the hovering monitoring platform will immediately send a preliminary alarm signal to the command center through the satellite communication unit, and record the specific time when the abnormal event occurred in the local storage module.

[0030] Step S140: Generate dynamic monitoring data according to the correlation between the environmental parameter data and the abnormal timestamp, where the dynamic monitoring data includes the time synchronization tag of the real-time image data and the spatial distribution information of the environmental parameter data.

[0031] Environmental parameter data is acquired in step S110, and an abnormal timestamp is marked in step S130. By analyzing the correlation between environmental parameter data and abnormal timestamps, we can understand the environmental conditions at the time of the abnormal event. For example, if an abnormal vehicle lane change occurs at a certain moment and the wind speed is high at that moment, this may indicate that the abnormal event is related to environmental factors.

[0032] Based on this correlation, dynamic monitoring data is generated. Time synchronization tags are added to real-time image data to ensure that the image data accurately corresponds to the time of the abnormal event. Simultaneously, environmental parameter data is processed to generate its spatial distribution information. For example, environmental parameter data collected by meteorological sensors at different locations is integrated to generate the spatial distribution of environmental parameters on the target road section. This generated dynamic monitoring data can more comprehensively and accurately reflect the actual conditions of the target road section, providing richer information for subsequent abnormal event identification and processing.

[0033] Step S200: identifying abnormal events existing in the target road section based on dynamic monitoring data, and generating multi-level warning signals according to the event type of the abnormal event.

[0034] Dynamic monitoring data includes time-synchronized tags for real-time image data and the spatial distribution of environmental parameter data. By analyzing and processing this data, abnormal events can be identified on target road sections. Abnormal event types include accidents, congestion, and obstacles. Multi-level warning signals are graded based on the severity and impact of abnormal events, enabling the command center and relevant personnel to take appropriate measures.

[0035] In one embodiment, in step S200, identifying abnormal events existing in the target road section based on dynamic monitoring data may specifically include the following steps S210 to S250:

[0036] Step S210: extracting vehicle density features, vehicle speed deviation features, and lane occupancy features from the dynamic monitoring data.

[0037] The vehicle density feature reflects the density of vehicles on the target road section, the speed deviation feature reflects the degree to which the vehicle speed differs from the normal speed, and the lane occupancy feature describes the lane usage. By extracting these features, abnormal events can be identified more accurately.

[0038] In one embodiment, step S210, extracting vehicle density features, vehicle speed deviation features, and lane occupancy features from dynamic monitoring data, may specifically include the following steps S211 to S218:

[0039] Step S211: Based on the time synchronization tag of the dynamic monitoring data, time alignment processing is performed on the real-time image data to obtain a synchronous monitoring data set that matches the spatial distribution information.

[0040] The time synchronization tags in the dynamic monitoring data provide accurate time information for the real-time image data. These time synchronization tags are used to perform time alignment processing on the real-time image data. The specific implementation method can be to sort the real-time image data collected at different times in chronological order, and then associate the image data at the same time and the same position according to the spatial distribution information to form a synchronous monitoring data set. The synchronous monitoring data set obtained in this way can more accurately reflect the actual situation of the target section at the same time, and provide a more reliable data basis for subsequent feature extraction. For example, for a certain section of highway, multiple real-time image data were collected at different time points. These image data were arranged in chronological order through time synchronization tags and associated with the corresponding spatial positions to obtain a synchronous monitoring data set that matches the spatial distribution information.

[0041] Step S212: Divide the target road section in the synchronous monitoring data set into a plurality of equally spaced monitoring areas, and count the frequency of vehicle appearances in each monitoring area within a preset time window to generate an initial density distribution set for each monitoring area.

[0042] The target road section is divided into multiple equally spaced monitoring areas, and the size of each monitoring area can be set according to actual conditions. The preset time window is a fixed time period, such as 1 minute or 5 minutes. Within this preset time window, the frequency of vehicle appearance in each monitoring area is counted. By counting the frequency of vehicle appearance in all monitoring areas, an initial density distribution set for each monitoring area is generated. This initial density distribution set reflects the preliminary distribution of vehicles on the target road section within the preset time window. For example, a section of highway is divided into 10 equally spaced monitoring areas. Within a preset time window of 5 minutes, the number of vehicle appearances in each monitoring area is counted to obtain the frequency of vehicle appearance in each monitoring area, thereby generating an initial density distribution set for each monitoring area.

[0043] Step S213: Perform spatial continuity analysis on the initial density distribution set, merge the areas in adjacent monitoring areas where the frequency of vehicle appearance exceeds the merging threshold, form a revised density distribution set, and generate vehicle density features based on the ratio of the total number of vehicles in each merged area to the area of ​​the area in the revised density distribution set.

[0044] Spatial continuity analysis analyzes the relationships between adjacent monitoring areas within the initial density distribution set. The merge threshold is a pre-set value used to determine whether adjacent monitoring areas should be merged. When the frequency of vehicle appearances in adjacent monitoring areas exceeds the merge threshold, it indicates strong continuity in the vehicle distribution between the two areas, and they are merged into a single area. This results in a revised density distribution set.

[0045] Next, the ratio of the total number of vehicles in each merged region in the revised density distribution set to the region's area is calculated. This ratio is the vehicle density feature. The vehicle density feature more accurately reflects the actual density of vehicles on the target road section. For example, in the initial density distribution set for a certain section of highway, the vehicle occurrence frequency in two adjacent monitoring regions, A and B, exceeds the merging threshold, so they are merged into one region, C. The total number of vehicles in region C is counted and the ratio of this number to the area of ​​region C is calculated to obtain the vehicle density feature for that region.

[0046] Step S214: extracting the real-time speed data of each vehicle in the target road section according to the environmental parameter data in the dynamic monitoring data, and calculating the absolute value of the speed difference between adjacent vehicles based on the real-time speed data.

[0047] The environmental parameter data in dynamic monitoring data contains relevant information about vehicles. By analyzing and processing this data, the real-time speed data of each vehicle on the target road section can be extracted. The real-time speed of the vehicle can be obtained using the speed sensors in the multi-directional sensor array, or by analyzing the real-time image data and using image processing algorithms to calculate the vehicle's driving speed.

[0048] After obtaining the real-time speed data for each vehicle, the absolute speed difference between adjacent vehicles is calculated. Specifically, for two adjacent vehicles, the speed of one vehicle is subtracted from the speed of the other, and the absolute value is taken. By calculating the absolute speed difference between all adjacent vehicles, the distribution of vehicle speed differences within the target road section can be determined.

[0049] Step S215: Sort the absolute values ​​of speed differences by vehicle travel direction, filter out abnormal speed segments where the absolute values ​​of speed differences of multiple consecutive vehicles exceed the fluctuation threshold, and count the distribution ratio of abnormal speed segments in the target road section to generate a vehicle speed deviation feature.

[0050] The absolute speed difference values ​​calculated in step S214 are sorted by vehicle travel direction, so that the absolute speed difference values ​​for vehicles traveling in the same direction are arranged sequentially. The fluctuation threshold is a pre-set value used to determine whether the vehicle speed difference is abnormal. When the absolute speed difference values ​​of multiple consecutive vehicles exceed the fluctuation threshold, the road section containing these vehicles is defined as an abnormal speed section.

[0051] The distribution ratio of abnormal speed segments within the target road section is calculated, and this ratio is the speed deviation feature. The speed deviation feature reflects the degree of vehicle speed instability within the target road section. For example, on a certain section of highway, after sorting the absolute values ​​of speed differences by vehicle travel direction, a continuous section of absolute speed differences exceeding the fluctuation threshold is found. This section is considered an abnormal speed segment. The ratio of the length of this abnormal speed segment to the entire target road section is calculated to obtain the speed deviation feature.

[0052] Step S216: Based on the lane identification data in the synchronous monitoring data set, identify the boundary line of each lane, and detect the lateral offset between the vehicle and the boundary line in the real-time image data.

[0053] The lane marking data in the synchronous monitoring data set contains the location and boundary information of each lane. By analyzing and processing this data, the lane boundaries can be identified. Image processing algorithms, such as edge detection algorithms, can be used to process real-time image data and extract lane boundaries.

[0054] After identifying the lane boundaries, the lateral offset between the vehicle and the boundary is detected in the real-time image data. This can be achieved by marking the vehicle's position and the lane boundary in the real-time image data, then calculating the lateral distance between the vehicle and the boundary. This distance is the lateral offset. For example, in real-time image data of a section of highway, the lane boundaries can be identified using an image processing algorithm. The position of each vehicle is then marked, and the lateral distance between the vehicle and the boundary is calculated to obtain the lateral offset.

[0055] Step S217: If the lateral offset exceeds the preset ratio of the lane width, the vehicle is marked as a lane-shifting vehicle, and the cumulative number of lane-shifting vehicle offsets per unit time in each lane is counted.

[0056] The lane width ratio is a pre-set value used to determine whether a vehicle has drifted from its lane. When the lateral offset between the vehicle and the boundary line exceeds the lane width ratio, the vehicle is marked as having drifted from its lane.

[0057] The cumulative number of lane deviations by vehicles in each lane is counted within a unit time. The unit time can be set based on actual conditions, such as 1 minute or 5 minutes. By counting the cumulative number of deviations, the lane deviation situation in each lane can be understood. For example, in a lane on a certain highway, within a unit time of 5 minutes, the lateral deviation of three vehicles exceeds a preset lane width ratio. These three vehicles are marked as lane deviation vehicles, and their cumulative number of deviations is counted.

[0058] Step S218: Determine the occupancy intensity coefficient of each lane based on the correlation between the cumulative number of offsets and the vehicle density characteristics, and mark the lanes whose occupancy intensity coefficient exceeds the warning coefficient as abnormally occupied lanes to generate lane occupancy characteristics.

[0059] Analyze the correlation between the cumulative number of lane deviations and vehicle density characteristics. A higher vehicle density indicates a higher probability of cumulative lane deviations, which can lead to more complex lane occupancy. A correlation model can be established to determine the occupancy intensity coefficient for each lane. This correlation model can be a linear regression model that takes the cumulative number of lane deviations and vehicle density characteristics as input and outputs the occupancy intensity coefficient for each lane.

[0060] The warning coefficient is a pre-set value used to determine whether a lane is abnormally occupied. When the occupancy intensity coefficient of a lane exceeds the warning coefficient, the lane is marked as abnormally occupied. By marking abnormally occupied lanes, a lane occupancy signature is generated. The lane occupancy signature reflects whether lane usage in the target road section is normal. For example, for a lane on a certain highway section, the occupancy intensity coefficient of the lane is calculated using an association model. When this coefficient exceeds the warning coefficient, the lane is marked as abnormally occupied, thus generating a lane occupancy signature.

[0061] Step S220: Match the vehicle density feature with the peak density in the historical traffic database to determine the first congestion index of the target road section.

[0062] The historical traffic database stores traffic data for the target section at different time periods, including information such as vehicle density. Peak density is the maximum vehicle density of the section recorded in the historical traffic database. The vehicle density feature generated in step S213 is matched with the peak density in the historical traffic database, and the first congestion index of the target section is determined by calculating the ratio of the vehicle density feature to the peak density. The first congestion index reflects the relative relationship between the current congestion level of the target section and the historical maximum congestion level. For example, if the vehicle density feature of a certain section of highway is currently 50 vehicles per square kilometer and the peak density in the historical traffic database is 100 vehicles per square kilometer, then the first congestion index of the section is 50 / 100=0.5.

[0063] Step S230: Calculate the speed difference gradient of adjacent vehicles based on the vehicle speed deviation characteristic. If the speed difference gradient exceeds a preset gradient threshold, generate a second congestion index.

[0064] The speed deviation characteristic reflects the degree of vehicle speed instability on the target road section. The speed difference gradient between adjacent vehicles is calculated based on the speed deviation characteristic. The speed difference gradient represents the rate of change of the speed difference between adjacent vehicles. Specifically, the speed difference gradient can be calculated by taking the absolute values ​​of the speed differences between two adjacent groups of vehicles, calculating their difference, and dividing it by the distance between them.

[0065] The preset gradient threshold is a pre-set value used to determine whether the change in vehicle speed difference is abnormal. When the speed difference gradient exceeds the preset gradient threshold, it indicates that the change in vehicle speed in the target section is large and congestion may exist. At this time, a second congestion index is generated based on the magnitude of the speed difference gradient. The second congestion index reflects the impact of vehicle speed changes in the target section on the degree of congestion. For example, on a certain section of highway, the speed difference gradient between adjacent vehicles is calculated to be 10km / h / km, and the preset gradient threshold is 5km / h / km. Since the speed difference gradient exceeds the gradient threshold, a second congestion index is generated based on this gradient value.

[0066] Step S240: Analyze the occupancy status of the emergency lane based on the lane occupancy characteristics. If the occupancy time of the emergency lane exceeds a preset response time limit, generate a third congestion index.

[0067] Lane occupancy characteristics contain information about the usage of each lane. By analyzing lane occupancy characteristics, we can understand the occupancy status of the emergency lane. The preset response time limit is a pre-set time value used to determine whether the emergency lane is occupied properly. If the emergency lane is occupied for longer than the preset response time limit, it indicates that the emergency lane is occupied unreasonably, which may affect emergency rescue operations and may also cause traffic congestion.

[0068] At this point, a third congestion index is generated based on the emergency lane's occupancy. This index reflects the impact of emergency lane occupancy on road congestion. For example, on a certain section of highway, if lane occupancy analysis reveals that the emergency lane is occupied, and the duration of the occupancy exceeds a preset response time limit, the third congestion index is generated based on this occupancy.

[0069] Step S250: Input the first congestion index, the second congestion index, and the third congestion index into an event classification model, and output an event type of the abnormal event, where the event type includes at least one of an accident type, a congestion type, and an obstacle type.

[0070] The event classification model can employ a machine learning model, such as a support vector machine (SVM). A support vector machine is a binary or multi-classification model that classifies data by finding the optimal classification hyperplane. In this patent, the first, second, and third congestion indices are used as input data to train a support vector machine model, enabling it to determine the type of abnormal event based on this input data.

[0071] When training a support vector machine model, we first collect a large amount of historical data, including the first, second, and third congestion indices under different circumstances, as well as the corresponding abnormal event types. This data is divided into a training set and a validation set. The support vector machine model is trained using the training set, and the model parameters are adjusted to enable the model to accurately classify abnormal event types. The trained model is then validated on the validation set to evaluate its performance.

[0072] In actual applications, the calculated first, second, and third congestion indices are input into a trained event classification model, which then outputs the type of abnormal event, such as accident type, congestion type, or obstacle type. For example, when the three congestion indices are input as 0.6, 0.3, and 0.2, respectively, the event classification model determines the abnormal event type as congestion type based on these data.

[0073] In one embodiment, step S200 generates a multi-level warning signal according to the event type of the abnormal event, which may specifically include the following steps S260 to S2120:

[0074] Step S260: Determine the target coverage range of the warning signal based on the event type. The target coverage range includes the lane where the abnormal event is located, adjacent lanes, and associated road sections within a preset distance in the target direction.

[0075] Different event types have varying impacts on traffic, so the target coverage range for warning signals needs to be determined based on the event type. For example, for an accident-related abnormal event, the target coverage range may include the lane where the accident occurred, adjacent lanes, and related road sections within a certain distance in front of and behind the accident site. For a congestion-related abnormal event, the target coverage range may be wider, encompassing the congested road section and its surrounding lanes and sections. The preset distance can be set based on actual conditions, such as 1 km or 2 km. By determining the target coverage range, we ensure that warning signals are accurately transmitted to potentially affected vehicles and personnel.

[0076] Step S270: Generate dynamic warning radiation parameters based on the correlation between vehicle density characteristics and target coverage range. The dynamic warning radiation parameters include signal propagation direction, signal strength attenuation gradient, and signal priority weight.

[0077] The vehicle density characteristic reflects the density of vehicles on the target road section, while the target coverage area determines the area that the warning signal needs to cover. Analyzing the correlation between the vehicle density characteristic and the target coverage area reveals that higher vehicle density may require stronger signal strength and more appropriate signal propagation direction to ensure that the warning signal is received by more vehicles.

[0078] Based on this correlation, dynamic warning radiation parameters are generated. The signal propagation direction is determined based on the target coverage range and the vehicle's direction of travel, ensuring that the warning signal propagates toward potentially affected vehicles. The signal strength attenuation gradient is adjusted based on vehicle density and distance. In areas with high vehicle density or long distances, the signal strength attenuation gradient can be set to a smaller value to ensure that the signal maintains a certain strength during propagation. The signal priority weight is set based on the severity of the abnormal event, and for more serious abnormal events, the signal priority weight can be set higher. For example, on a certain section of highway with high vehicle density, the target coverage range includes the 2-kilometer section ahead. Based on this information, dynamic warning radiation parameters are generated, with the signal propagation direction set to the front, the signal strength attenuation gradient set to a smaller value, and the signal priority weight set to a higher value based on the severity of the abnormal event.

[0079] Step S280: Send a pre-carrier detection signal to vehicles within the target coverage area through the signal relay module of the hovering monitoring platform, and receive a carrier response signal fed back by the vehicle.

[0080] The hovering monitoring platform's signal relay module is responsible for signal relay and transmission. It sends a pre-carrier detection signal to vehicles within the target coverage area. The pre-carrier detection signal is used to detect a vehicle's signal reception capability. Upon receiving the pre-carrier detection signal, the vehicle returns a carrier response signal. By receiving the carrier response signal from the vehicle, we can understand the vehicle's reception status and provide a basis for subsequent adjustments to the warning signal parameters. For example, the hovering monitoring platform sends a pre-carrier detection signal to vehicles within the target coverage area. After receiving the signal, some vehicles will return a carrier response signal via their onboard equipment, which the hovering monitoring platform then receives.

[0081] Step S290: adjusting the signal strength attenuation gradient in the dynamic warning radiation parameter according to the signal quality parameter of the carrier response signal, and generating a directional enhancement signal based on the adjusted signal strength attenuation gradient.

[0082] Signal quality parameters for the carrier response signal include signal strength and signal-to-noise ratio. When analyzing these signal quality parameters, if the signal quality is poor, it indicates that the signal strength received by the vehicle may be insufficient, and the signal strength attenuation gradient in the dynamic warning radiation parameters needs to be adjusted. Specifically, this adjustment method involves comparing the signal quality parameters with preset standard values. If the signal quality falls below the standard, the signal strength attenuation gradient is reduced to increase signal strength during propagation.

[0083] Based on the adjusted signal strength attenuation gradient, a directional enhancement signal is generated. This directional enhancement signal is a warning signal with stronger signal strength and more accurate propagation direction, enabling better transmission to vehicles within the target coverage area. For example, by analyzing the signal quality parameters of the carrier response signal, it is found that the signal strength received by some vehicles is weak. In this way, the signal strength attenuation gradient is reduced, and then a directional enhancement signal is generated based on the adjusted signal strength attenuation gradient.

[0084] Step S2100: Encapsulate the event type, target coverage, and directional enhancement signal into a multi-level warning signal, and spatially associate the multi-level warning signal with the geographic location tag of the hovering monitoring platform to generate an encrypted data packet.

[0085] The event type, target coverage, and directional enhancement signal are packaged to form a multi-level warning signal. This multi-level warning signal contains detailed information about the abnormal event and the warning content. The multi-level warning signal is then spatially and temporally correlated with the geolocation tag of the hovering monitoring platform, which records the specific location and time of the hovering monitoring platform. This spatial and temporal correlation allows the precise location and time of the abnormal event.

[0086] Finally, the encapsulated and correlated information is encrypted to generate an encrypted data packet. Symmetric encryption algorithms, such as AES, can be used for encryption to ensure data packet security during transmission. For example, an abnormal event, corresponding target coverage, and directional enhancement signals can be encapsulated into a multi-level warning signal. This signal is then spatially and temporally correlated with the geolocation tag of the hovering monitoring platform and encrypted using the AES algorithm to generate an encrypted data packet.

[0087] Step S2110: Send the encrypted data packet to the command center via the satellite communication unit, and monitor the transmission integrity of the encrypted data packet in real time during the sending process.

[0088] The satellite communication unit is responsible for sending encrypted data packets to the command center via satellite. During the transmission process, the transmission integrity of the encrypted data packets is monitored in real time. A checksum algorithm, such as the CRC algorithm, can be used to verify the encrypted data packets. At the sending end, the checksum value of the encrypted data packet is calculated and appended to the data packet; at the receiving end, the checksum value of the received data packet is recalculated and compared with the appended checksum value. If the two checksum values ​​are the same, the data packet transmission is complete; if they are different, it means that an error may have occurred during the data packet transmission process. For example, the hovering monitoring platform sends encrypted data packets to the command center via the satellite communication unit, while using the CRC algorithm to calculate the checksum value of the data packet. Changes in the checksum value are monitored in real time during the transmission process to ensure the transmission integrity of the data packet.

[0089] Step S2120: If the transmission integrity is detected to be lower than the preset integrity threshold, the local storage module of the hovering monitoring platform is triggered to temporarily store the multi-level warning signal and retransmit the encrypted data packet through the backup communication link until the command center returns a reception confirmation signal.

[0090] The preset integrity threshold is a pre-set value used to determine the integrity of data packet transmission. If the transmission integrity is detected below the preset integrity threshold, it indicates that the data packet may have been lost or damaged during transmission. At this time, the local storage module of the hovering monitoring platform is triggered to temporarily store multi-level warning signals to prevent data loss.

[0091] At the same time, the encrypted data packet is retransmitted via a backup communication link. The backup communication link can be another communication method, such as a ground communication network. The encrypted data packet is continuously retransmitted until the command center returns a receipt confirmation signal, indicating that the command center has successfully received the encrypted data packet. For example, if, during the transmission of the encrypted data packet, the transmission integrity is detected to be below a preset integrity threshold, the local storage module of the hovering monitoring platform temporarily stores the multi-level warning signal and then retransmits the encrypted data packet via the ground communication network until the command center returns a receipt confirmation signal.

[0092] Step S300: Transmit the multi-level warning signal to the command center, triggering the command center to generate a drone dispatch instruction that matches the event type. The drone dispatch instruction includes the identification of the target drone and path planning parameters.

[0093] The multi-level warning signal contains detailed information about the abnormal event and its type. After transmitting the multi-level warning signal to the command center, the command center generates a matching drone dispatch instruction based on the event type. The target drone's identifier uniquely identifies the drone to be dispatched, while the path planning parameters define the drone's flight path and related parameters.

[0094] In one embodiment, in step S300, triggering the command center to generate a drone dispatch instruction that matches the event type may specifically include the following steps S310 to S370:

[0095] Step S310: Based on the event type and target coverage in the multi-level warning signal, a set of candidate drones with corresponding functional modules is screened from the drone resource pool.

[0096] The drone resource pool contains multiple drones with varying functions and performance. Based on the event type and target coverage in the multi-level warning signal, drones with corresponding functional modules are selected as candidate drones. For example, for an accident-related abnormal event, drones with image acquisition modules and audio and visual warning modules are selected; for a congestion-related abnormal event, drones with voice broadcast modules and communication modules are selected. By screening the candidate drone set, we can ensure that the dispatched drones can meet the needs of abnormal event handling.

[0097] Step S320: Calculate the scheduling priority score of each UAV in the candidate UAV set based on its endurance parameters, current mission load status, and relative distance to the target location.

[0098] The endurance parameter reflects the time and distance a drone can sustain flight. The current mission load status indicates whether the drone is currently performing other tasks. The relative distance to the target location indicates the distance between the drone and the location of the abnormal event. Based on these three factors, each drone's scheduling priority score is calculated. A weighted summation method can be used to assign different weights to the endurance parameter, current mission load status, and relative distance to the target location. The scheduling priority score is then calculated by multiplying the values ​​of each factor by the corresponding weight and adding them together.

[0099] Step S330: Sort the candidate drone set according to the scheduling priority score, and select the drone with the highest score as the target drone.

[0100] The candidate drones are sorted from high to low according to their scheduling priority scores. The drone with the highest score is the most suitable for the current task. The drone with the highest score is selected as the target drone to ensure that abnormal events can be handled efficiently.

[0101] Step S340: Extract the signal propagation direction and signal strength attenuation gradient from the dynamic warning radiation parameters to generate the initial path planning parameters for the target UAV.

[0102] The signal propagation direction and signal strength attenuation gradient in the dynamic warning radiation parameters reflect the impact range and signal propagation of the abnormal event. These parameters are extracted to determine the target drone's flight direction based on the signal propagation direction. The drone's flight altitude and speed are adjusted based on the signal strength attenuation gradient to generate the target drone's initial path planning parameters. These initial path planning parameters provide preliminary guidance for the drone's flight. For example, based on the signal propagation direction, the target drone is directed to fly toward the location of the abnormal event. Based on the signal strength attenuation gradient, the drone's flight altitude is adjusted to 50 meters and its speed to 30 kilometers per hour to generate the target drone's initial path planning parameters.

[0103] Step S350: Based on the real-time environmental parameter data of the target location, the initial path planning parameters are dynamically modified to obtain path planning parameters including obstacle avoidance path nodes, flight altitude gradient, and return charging strategy.

[0104] Real-time environmental parameter data at the target location includes information such as weather conditions and topography. This information may affect the drone's flight safety and efficiency. Based on this real-time environmental parameter data, the initial path planning parameters are dynamically adjusted. For example, if the wind speed at the target location is high, the drone's flight altitude and speed may need to be adjusted; if there are obstacles at the target location, obstacle avoidance path nodes may need to be planned.

[0105] Through dynamic correction, the path planning parameters, including obstacle avoidance path nodes, flight altitude gradients, and return-to-home charging strategies, are obtained. The obstacle avoidance path nodes specify the specific locations and routes for avoiding obstacles during flight; the flight altitude gradient represents the change in the drone's flight altitude at different locations; and the return-to-home charging strategy determines the return route and charging method for the drone when the battery is low. For example, based on real-time environmental parameter data at the target location, if a tall building is detected ahead, an obstacle avoidance path node is added to the path planning parameters to allow the drone to fly around the building. Simultaneously, the flight altitude gradient and return-to-home charging strategy are adjusted based on wind speed and battery conditions.

[0106] Step S360: Encapsulate the target UAV's identification, path planning parameters, and event processing requirements of the multi-level warning signal into instructions to generate a UAV dispatch instruction.

[0107] The target drone's identification, path planning parameters, and event handling requirements for the multi-level warning signal are encapsulated into a complete instruction package, known as the drone dispatch instruction. This instruction package contains all the information the drone needs to execute its mission, and the command center can use it to convey the mission requirements to the target drone. For example, the target drone is identified as "Drone A," and the path planning parameters include obstacle avoidance path nodes, flight altitude gradient, and return-to-home charging strategy. The event handling requirements for the multi-level warning signal include image capture and audio and visual warnings at the accident scene. This information is encapsulated into the drone dispatch instruction.

[0108] Step S370: Provide a pre-charging interface for the target UAV through the energy supply module of the hovering monitoring platform, and verify its compatibility with the path planning parameters before the target UAV takes off.

[0109] The hovering monitoring platform's energy supply module provides a pre-charging interface for the target drone, ensuring it has sufficient power before takeoff. Before the target drone takes off, its compatibility with the path planning parameters is verified. The drone's flight performance parameters can be checked to ensure they meet the requirements of the path planning parameters. For example, the drone's maximum flight speed can reach the speed specified in the path planning parameters, and its endurance can support flight to the target location and return according to the path planning parameters. If incompatibility exists, the path planning parameters must be adjusted or a suitable drone must be replaced. For example, if the target drone's maximum flight speed is 40 kilometers per hour, while the path planning parameters specify a flight speed of 50 kilometers per hour, this indicates an incompatibility and the flight speed in the path planning parameters needs to be adjusted.

[0110] Step S400: Control the target UAV to move to the target location corresponding to the abnormal event based on the path planning parameters, and perform an early warning operation corresponding to the event type through the external speaker module of the target UAV.

[0111] The target drone moves to the target location corresponding to the abnormal event based on the flight route, altitude, and speed information in the path planning parameters. Upon arrival, the target drone's external speaker module performs a warning operation corresponding to the event type, alerting surrounding vehicles and personnel to the abnormal event.

[0112] In one embodiment, in step S400, controlling the target UAV to move to the target location corresponding to the abnormal event based on the path planning parameters may specifically include the following steps S410 to S470:

[0113] Step S410: After the target UAV takes off, a navigation calibration signal is continuously sent to the target UAV via the signal relay module of the hovering monitoring platform. The navigation calibration signal includes a real-time updated target position offset and environmental interference compensation parameters.

[0114] The signal relay module of the hovering monitoring platform is responsible for sending navigation calibration signals to the target drone. After the target drone takes off, the target position may shift due to environmental factors, and environmental interference may also affect the drone's flight. Therefore, it is necessary to update the target position offset and environmental interference compensation parameters in real time, and include this information in the navigation calibration signal sent to the target drone. The target position offset represents the change in the target position relative to the initial position, and the environmental interference compensation parameters are used to compensate for the impact of environmental interference on the drone's flight. For example, due to changes in wind direction, the target position may shift to a certain extent. The hovering monitoring platform monitors the changes in the target position in real time, calculates the target position offset, and includes it in the navigation calibration signal along with the environmental interference compensation parameters and sends it to the target drone.

[0115] Step S420: adjusting the flight altitude gradient of the target UAV based on the environmental interference compensation parameter, and collecting the attitude data and energy consumption data of the target UAV in real time.

[0116] The target drone's flight altitude gradient is adjusted based on the environmental interference compensation parameters in the navigation calibration signal. When environmental interference is high, the altitude gradient is adjusted appropriately to ensure the drone's flight safety and stability. Simultaneously, attitude data and energy consumption data are collected from the target drone in real time. Attitude data, including pitch, roll, and yaw angles, reflects the drone's flight attitude; energy consumption data indicates the drone's power consumption during flight. Real-time data collection provides timely information on the drone's flight status and battery life. For example, when the environmental interference compensation parameters indicate high wind speeds, the target drone's flight altitude gradient is adjusted to appropriately increase its flight altitude. Simultaneously, sensors collect the drone's attitude and energy consumption data in real time.

[0117] Step S430: Compare the posture data with the standard posture range in the path planning parameters. If it is detected that the posture deviation exceeds the fault tolerance threshold, a dynamic obstacle avoidance path is generated according to the obstacle avoidance path node.

[0118] The path planning parameters specify the target drone's standard attitude range. Real-time attitude data is compared against this standard range. A tolerance threshold is a pre-set value used to determine whether attitude deviations are within an acceptable range. If the detected attitude deviation exceeds the tolerance threshold, it indicates a significant anomaly in the drone's flight attitude, potentially compromising flight safety. In this case, a dynamic obstacle avoidance path is generated based on the obstacle avoidance path nodes in the path planning parameters. These nodes contain pre-planned positions and routes for avoiding obstacles. By analyzing and processing these nodes and combining them with the drone's current position and attitude, a dynamic obstacle avoidance path is generated, allowing the drone to safely avoid obstacles and continue flight. For example, if the standard pitch angle range specified in the path planning parameters is -10° to 10°, and real-time attitude data indicates a pitch angle of 15°, exceeding the tolerance threshold of 5°, a dynamic obstacle avoidance path is generated based on the obstacle avoidance path nodes, guiding the drone to change flight direction and avoid potential hazards.

[0119] Step S440: Calculate the remaining flight time of the target UAV based on the energy consumption data, and adjust the charging priority in the return charging strategy based on the remaining flight time.

[0120] Energy consumption data reflects the target drone's power consumption during flight. By analyzing and compiling this data, the target drone's remaining flight time can be calculated. This calculation method involves calculating the power consumption per unit time based on the consumed power and flight time. The remaining power is then divided by the power consumption per unit time to determine the remaining flight time.

[0121] Adjust the charging priority in the return-to-home charging strategy based on the remaining flight time. When the remaining flight time is short, increase the charging priority to ensure the drone can return to recharge in time to avoid mission failure or danger due to battery depletion. For example, if the target drone's remaining flight time is calculated to be 20 minutes, but it will take another 15 minutes to reach the target location of the abnormal event, and the completion of the warning operation may also require some time, increase the charging priority and arrange for the drone to return to the nearest hovering monitoring platform for charging as soon as possible after the warning operation is completed.

[0122] Step S450: When the target UAV approaches the target location, the real-time distance data to the target location is obtained through the positioning module carried by the target UAV.

[0123] The target drone's onboard positioning module can be a Global Positioning System (GPS) or other high-precision positioning device. As the drone approaches the target location, it uses the positioning module to obtain real-time distance data to the target. By receiving satellite signals or other positioning signals, the positioning module accurately calculates the drone's current coordinates, then compares them with the coordinates of the target location to calculate the distance between them. For example, while the target drone is in flight, the positioning module continuously receives GPS signals. When it approaches the target location of an abnormal event, it calculates and obtains the distance to the target location in real time, such as displaying a distance of 500 meters.

[0124] Step S460: If the real-time distance data is less than the preset trigger threshold, the external speaker module of the target UAV is activated to enter the warning ready state, and the image acquisition module is activated simultaneously.

[0125] The preset trigger threshold is a pre-set distance value used to determine whether the target drone is close enough to initiate an early warning operation. When the real-time distance data is less than the preset trigger threshold, it means that the drone is close enough to the target location and the early warning operation can be started. At this time, the external speaker module of the target drone is activated to enter the early warning ready state. The external speaker module may include speakers, lights and other equipment for emitting sound and light warning signals. At the same time, the image acquisition module is synchronously activated. The image acquisition module can be a high-definition camera for collecting image data of the scene of the abnormal event. For example, the preset trigger threshold is 200 meters. When the real-time distance data of the target drone shows 150 meters, the external speaker module and the image acquisition module are activated.

[0126] Step S470: The real-time status data of the target UAV is transmitted back to the command center through the signal relay module, triggering the command center to update the path planning parameters in the UAV scheduling instruction.

[0127] The target drone's real-time status data includes information such as its attitude, energy consumption, and real-time distance to the target location. This real-time status data is transmitted back to the command center via the hovering monitoring platform's signal relay module. Upon receiving this data, the command center assesses the target drone's current status. If any changes in the target drone's status are detected, such as a deviation in attitude or rapid battery depletion, the command center will update the path planning parameters in the drone's dispatch instructions accordingly. For example, if the command center discovers that the target drone's battery is depleting faster than expected based on the real-time status data transmitted back, it will update the path planning parameters and adjust the drone's flight speed and route to ensure it returns to recharge as quickly as possible.

[0128] In one embodiment, in step S400, the warning operation corresponding to the event type is executed by the external speaker module of the target drone, which may specifically include the following steps S480 to S4140:

[0129] Step S480: Calling the preset warning mode library of the external speaker module according to the event type, the warning mode library includes sound and light warning combinations, voice broadcast templates and laser positioning modes.

[0130] The external speaker module's preset warning mode library contains a collection of pre-set warning modes for different event types. Based on the abnormal event type determined in step S250, the corresponding warning mode is invoked from the preset warning mode library. For example, for an accident-related abnormal event, a combination of audio and visual warnings and image acquisition might be invoked; for a congestion-related abnormal event, a voice announcement template might be invoked and a deceleration control command might be sent to adjacent vehicles; for an obstacle-related abnormal event, a laser positioning mode might be invoked and three-dimensional obstacle contour data generated. By invoking the preset warning mode library, the warning action corresponding to the event type can be quickly and accurately executed.

[0131] Step S490: If the event type is an accident type, the directional sound wave pulse and strobe light sequence in the sound and light warning combination are selected, and the image acquisition module is simultaneously started to record a panoramic video stream of the accident scene.

[0132] When the abnormal event type is an accident, a directional sound pulse and a strobe light sequence are selected from the sound and light warning combination. The directional sound pulse can transmit the warning sound to a specific area, attracting the attention of surrounding vehicles and personnel; the strobe light sequence uses rapidly flashing lights to visually alert surrounding traffic participants. Simultaneously, the image acquisition module is activated. This module can be a high-definition camera with wide-angle capabilities to record a panoramic video stream of the accident scene. The panoramic video stream records detailed information about the accident scene, including the location of the accident vehicle, damage, and the condition of the personnel, providing important information for subsequent accident handling and analysis. For example, if a vehicle collision occurs on a highway, upon arrival, the target drone activates a directional sound pulse to sound a warning sound, while the strobe light rapidly flashes. The image acquisition module begins recording a panoramic video stream of the accident scene.

[0133] Step S4100: If the event type is congestion type, select the multilingual alternative route prompt in the voice broadcast template, and send a deceleration control instruction to the vehicle-mounted terminal of the adjacent vehicle through the communication module of the target drone.

[0134] When the abnormal event type is congestion, select multilingual alternative route prompts from the voice broadcast template. Multilingual alternative route prompts can provide clear route guidance for drivers of different language backgrounds, helping them avoid congested sections. The communication module of the target drone can establish a communication connection with the on-board terminal of the adjacent vehicle and send deceleration control instructions to the on-board terminal through wireless communication technology. The deceleration control instruction can remind the driver to reduce the speed to avoid accidents such as rear-end collisions, and also help alleviate congestion. For example, when congestion occurs on a certain section of highway, the target drone prompts the driver to choose an alternative route through a voice broadcast template in multiple languages ​​such as Chinese and English, and at the same time sends deceleration control instructions to the on-board terminal of the adjacent vehicle, suggesting that the driver reduce the speed to 30 kilometers per hour.

[0135] Step S4110: If the event type is an obstacle type, activate the laser positioning mode to project the obstacle boundary mark and generate three-dimensional contour data of the obstacle.

[0136] When the abnormal event type is an obstacle, the laser positioning mode is activated. Laser positioning mode emits a laser beam, scans the surface of the obstacle, and projects the obstacle's boundary marker, allowing surrounding vehicles and personnel to clearly see the obstacle's location and range. Simultaneously, the laser scanning data is used to generate 3D contour data of the obstacle. This 3D contour data accurately describes the obstacle's shape, size, and location, providing a critical reference for subsequent obstacle removal and traffic diversion. For example, if a piece of fallen cargo is discovered on the highway as an obstacle, the target drone activates laser positioning mode, projects the cargo's boundary marker, and simultaneously generates 3D contour data of the cargo, displaying its length, width, height, and specific location.

[0137] Step S4120: Transmit the panoramic video stream, deceleration control instructions and three-dimensional contour data to the command center in real time, triggering the command center to update the event processing status of the multi-level warning signal according to the panoramic video stream.

[0138] The target drone's communication module transmits a panoramic video stream, deceleration control instructions, and 3D contour data to the command center in real time. After receiving this data, the command center evaluates the handling of the abnormal event based on the panoramic video stream. If changes are detected at the accident scene, such as the accident vehicle being towed away or congestion being alleviated, the command center updates the event handling status of the multi-level warning signal. The event handling status can range from handled, in progress, to unhandled. Timely and accurate updates help relevant departments and personnel understand the progress of the abnormal event and take appropriate measures. For example, if the command center detects that the vehicles at the accident scene have been largely cleared based on the panoramic video stream received, it will update the event handling status of the multi-level warning signal to handled.

[0139] Step S4130: When the event processing status changes to completed, the target UAV is controlled to return to the nearest hovering monitoring platform based on the adjusted return charging strategy.

[0140] When the command center updates the event handling status to Completed based on data such as the panoramic video stream, it indicates that the abnormal event has been properly handled and the target drone can conclude its mission and return to recharge. The target drone is then controlled to return to the nearest hovering monitoring platform based on the adjusted return-to-home charging strategy. This adjusted return-to-home charging strategy is adjusted during flight based on energy consumption data and actual conditions to ensure the drone's safe and efficient return to recharge. For example, after the target drone completes its warning and image collection mission at the accident scene, the command center updates the event handling status to Completed. The drone then follows the planned route back to the nearest hovering monitoring platform based on the adjusted return-to-home charging strategy.

[0141] Step S4140: During the return process, the remaining flight time of the target UAV is continuously monitored. If the remaining flight time is lower than the safety threshold, the emergency charging protocol is activated and mid-flight charging is performed through the hovering monitoring platform passing by.

[0142] During the target drone's return, its remaining flight time is continuously monitored. The safety threshold is a pre-set battery level. When the remaining flight time falls below the safety threshold, the drone's battery is insufficient to return to the hovering monitoring platform, posing a risk of battery depletion. At this point, the emergency charging protocol is activated. This protocol specifies how the drone communicates with and charges from the hovering monitoring platform it's passing through when its battery is low. The drone establishes a communication connection with the hovering monitoring platform and requests charging. The hovering monitoring platform's energy supply module provides the drone with a charging port for mid-flight charging. For example, if the target drone detects that its remaining flight time has fallen below the safety threshold during its return, it will immediately activate the emergency charging protocol, communicate with the nearest passing hovering monitoring platform, and dock there for mid-flight charging.

[0143] Step S500: Adjust the monitoring range of the hovering monitoring platform according to the persistence of the abnormal event, and send a flow adjustment instruction to the service area control system of the target direction vehicle to dynamically control the traffic strategy of the service area entrance.

[0144] The ongoing status of an abnormal event includes information such as whether the incident has been resolved and whether there is a trend of spread. Based on this information, the hovering monitoring platform adjusts its monitoring range to ensure timely detection of changes in abnormal events. Simultaneously, flow control instructions are sent to the service area control system for vehicles in the target direction. By adjusting the access strategy at the service area entrance, traffic flow is dynamically controlled to prevent further congestion.

[0145] In one embodiment, in step S500, adjusting the monitoring range of the hovering monitoring platform according to the persistence of the abnormal event may specifically include the following steps S510 to S570:

[0146] Step S510: Based on the event processing status fed back by the command center and the real-time status data of the target UAV, the diffusion rate of the abnormal event and the impact boundary change trend are calculated.

[0147] The incident handling status reported by the command center reflects the progress of handling the abnormal event. The target drone's real-time status data includes information such as the on-site images and environmental parameters it collects. By analyzing and processing this data, the abnormal event's diffusion rate and impact boundary change trend are calculated. The diffusion rate indicates how quickly the abnormal event spreads over time and space, while the impact boundary change trend describes the expansion or contraction of the abnormal event's impact range. For example, if the command center reports that a traffic accident is still under investigation, but the target drone's real-time status data shows that the length of the vehicle queue around the accident site is increasing, this indicates that the abnormal event is spreading. Further analysis can calculate the diffusion rate and the direction and speed of change of the impact boundary.

[0148] Step S520: If the diffusion rate exceeds the capture capability corresponding to the current monitoring coverage radius of the hovering monitoring platform, the collaborative monitoring protocol of the adjacent hovering monitoring platform is activated. The collaborative monitoring protocol includes data synchronization frequency adjustment rules and sensor array scanning angle complementary strategy.

[0149] The current monitoring coverage radius of a hovering monitoring platform is the range within which it can effectively monitor. When the diffusion rate of an abnormal event exceeds the capture capability corresponding to this coverage radius, a single hovering monitoring platform is unable to fully monitor changes in the abnormal event. At this point, the collaborative monitoring protocol of adjacent hovering monitoring platforms is activated. The data synchronization frequency adjustment rule specifies the frequency of data transmission between adjacent hovering monitoring platforms. By increasing the data synchronization frequency, timely data sharing between platforms is ensured. The sensor array scanning angle complementation strategy involves coordinating the scanning angles of the multi-directional sensor arrays of adjacent hovering monitoring platforms to avoid blind spots and expand the monitoring range. For example, if a hovering monitoring platform has a monitoring coverage radius of 1 kilometer and the diffusion rate of an abnormal event indicates that its impact range is expanding at a rate of 2 kilometers per hour, exceeding the platform's capture capability, the collaborative monitoring protocol of adjacent hovering monitoring platforms is activated, adjusting the data synchronization frequency to once per minute and ensuring that the scanning angles of the sensor arrays of adjacent platforms complement each other.

[0150] Step S530: The dynamic monitoring data of multiple hovering monitoring platforms are temporally and spatially aligned and fused through a collaborative monitoring protocol to generate an extended monitoring data set. The extended monitoring data set includes cross-platform vehicle trajectory continuity characteristics and environmental parameter superposition effect data.

[0151] After activating the collaborative monitoring protocol, data sharing begins between multiple hovering monitoring platforms. Through the data synchronization mechanism within the collaborative monitoring protocol, the dynamic monitoring data collected by each platform is spatially and temporally aligned and fused. Spatial and temporal alignment and fusion refers to matching and integrating data collected by different platforms at different times and spaces to uniformly define the data's temporal and spatial coordinates. The resulting extended monitoring data set includes cross-platform vehicle trajectory continuity features and environmental parameter overlay effect data. The cross-platform vehicle trajectory continuity feature tracks the complete driving trajectory of a vehicle within the monitoring range of multiple hovering monitoring platforms. The environmental parameter overlay effect data integrates environmental parameters collected by multiple platforms to more comprehensively reflect the environmental conditions in the area affected by the abnormal event. For example, two adjacent hovering monitoring platforms each collected vehicle driving information on different road sections. Through spatial and temporal alignment and fusion, the continuous driving trajectory of the vehicle on these two sections is obtained. Simultaneously, the meteorological parameters collected by the two platforms are overlaid and analyzed to obtain more accurate environmental parameter overlay effect data.

[0152] Step S540: recalculate the spatial distribution gradient of the vehicle density characteristics based on the extended monitoring data set, and identify new congestion nodes or accident-derived areas that affect the boundary change trend.

[0153] The extended monitoring data set provides more comprehensive and accurate traffic information. Based on this data, the spatial distribution gradient of vehicle density characteristics is recalculated. The spatial distribution gradient represents the change in vehicle density at different locations. By analyzing the spatial distribution gradient of vehicle density characteristics, it is possible to identify new congestion nodes or accident-derived areas that affect boundary change trends. New congestion nodes refer to newly emerged traffic congestion points under the influence of abnormal events, and accident-derived areas refer to traffic anomalies in surrounding areas caused by accidents. For example, through analysis of the extended monitoring data set, it was found that the spatial distribution gradient of vehicle density characteristics in a certain area suddenly increased, indicating that a new congestion node may have appeared in the area. Further analysis may show that vehicles are queuing in the area due to an accident.

[0154] Step S550: Based on the position coordinates of the newly added congested node or accident-derived area, the scanning focus direction of the multi-directional sensor array of the hovering monitoring platform is adjusted, and the image acquisition frame rate of the newly added congested node is increased.

[0155] After identifying newly added congested nodes or accident-derived areas, the scanning focus direction of the multi-directional sensor array of the hovering monitoring platform is adjusted according to its position coordinates. The scanning direction of the sensor array is focused on the newly added congested nodes or accident-derived areas to monitor the conditions in these areas in more detail. At the same time, the image acquisition frame rate of the newly added congested nodes is increased, and the frequency of image data acquisition is increased, so that changes in the area can be captured more promptly. For example, if a newly added congested node is found to be located at a specific location, the scanning direction of the multi-directional sensor array of the hovering monitoring platform is adjusted to that location, and the image acquisition frame rate is increased from 1 frame per second to 5 frames per second.

[0156] Step S560: Encapsulate the adjusted scanning focus direction and image acquisition frame rate parameters into a monitoring range update instruction, and send it to all hovering monitoring platforms participating in the collaborative monitoring protocol.

[0157] The adjusted scanning focus direction and image acquisition frame rate parameters are encapsulated to form a monitoring range update instruction. This instruction contains the monitoring parameter information that needs to be updated for each hovering monitoring platform. The monitoring range update instruction is sent to all hovering monitoring platforms participating in the collaborative monitoring protocol through the communication module of the hovering monitoring platform. After receiving the instruction, each platform updates its own monitoring settings according to the parameters in the instruction to ensure that the entire collaborative monitoring system can more effectively monitor changes in abnormal events. For example, the scanning focus direction is adjusted to focus on the direction of the newly added congested node, and the image acquisition frame rate is increased to 5 frames per second. The parameters are encapsulated as a monitoring range update instruction and sent to all hovering monitoring platforms participating in the collaborative monitoring protocol through the satellite communication unit.

[0158] Step S570: Receive feedback on the status of each platform executing the monitoring range update instruction in real time through the satellite communication unit. If it is detected that some platforms have not completed the parameter update, activate the backup hovering monitoring platform to replace the platform that has not responded to the instruction, and reallocate its monitoring coverage area.

[0159] The satellite communication unit is used to receive in real time the feedback status of the execution of the monitoring range update instruction by each hovering monitoring platform participating in the collaborative monitoring protocol. The feedback status may include information such as whether the parameter update is successful and the time of the update. If it is detected that some platforms have not completed the parameter update, it may be due to communication failure, equipment failure, etc. At this time, the backup hovering monitoring platform is activated to replace the platform that did not respond to the instruction. The backup hovering monitoring platform is pre-set and has the same monitoring functions and performance. After the replacement, its monitoring coverage area is reallocated to ensure that the monitoring range of the entire monitoring system is not affected. For example, if it is discovered through the satellite communication unit that a certain hovering monitoring platform has not completed the parameter update of the monitoring range update instruction, the backup hovering monitoring platform is immediately activated to replace the platform, and the monitoring coverage area of ​​the original platform is allocated to the backup platform.

[0160] In one embodiment, in step S500, sending a flow adjustment instruction to a service area control system of a vehicle in the target direction may specifically include the following steps S580 to S5140:

[0161] Step S580: Based on the vehicle trajectory continuity characteristics in the extended monitoring data set, the probability distribution of target direction vehicles arriving at the service area within a preset time period and the traffic peak period are predicted.

[0162] The vehicle trajectory continuity feature in the extended monitoring data set records vehicle travel trajectory and speed information. By analyzing and mining this information, combined with historical traffic data and real-time traffic conditions, the probability distribution of vehicles arriving at the service area in the target direction within a preset time period and the peak traffic period are predicted. The probability distribution represents the likelihood of vehicles arriving at the service area at different times, and the peak traffic period refers to the period with the highest vehicle flow within the preset time period. For example, based on the travel trajectory and speed of vehicles in the extended monitoring data set, the probability distribution of vehicles arriving at the service area in a certain direction within the next hour is analyzed, and the peak traffic period is predicted to occur between 30 and 45 minutes.

[0163] Step S590: Generate dynamic traffic control parameters for the service area entrance based on the probability distribution and traffic peak period. The dynamic traffic control parameters include the entrance lane allocation strategy, signal light phase switching rules, and variable information board content update cycle.

[0164] Based on the predicted probability distribution of vehicles arriving at the service area and peak traffic periods, dynamic flow control parameters are generated for the service area entrance. The entrance lane allocation strategy refers to how to rationally allocate lanes at the service area entrance to improve vehicle traffic efficiency. For example, during peak traffic periods, the number of lanes entering the service area can be increased. The signal light phase switching rule specifies the on / off timing and sequence of the entrance signal lights. Adjusting the signal light phase according to traffic conditions reduces vehicle waiting time. The variable message board content update cycle refers to the frequency of updating the information displayed on the variable message board. This information is updated promptly based on real-time traffic conditions to provide accurate guidance for drivers. For example, based on the probability distribution and peak traffic periods, a lane allocation strategy is developed to increase the number of entrance lanes from two to three during peak periods, adjust the signal light phase switching rule to extend the green light time, and adjust the variable message board content update cycle to every five minutes.

[0165] Step S5100: Obtain real-time entrance traffic flow data and parking space occupancy status through the service area control system, and perform matching analysis with dynamic traffic control parameters.

[0166] The service area control system monitors traffic flow data and parking space occupancy status at the service area entrance in real time. By performing a match analysis with the dynamic traffic control parameters, the rationality and effectiveness of the dynamic traffic control parameters are evaluated. If the real-time entrance traffic flow data and parking space occupancy status do not match the dynamic traffic control parameters, the current control parameters may need to be adjusted. For example, if the dynamic traffic control parameters predict that the entrance traffic volume will be high during a certain time period, requiring an increase in the number of lanes, but the actual real-time entrance traffic flow data shows low traffic volume, the lane allocation strategy may need to be adjusted.

[0167] Step S5110: If the deviation between the entrance traffic flow data and the predicted traffic peak period exceeds the preset tolerance range, the priority vehicle type and release interval duration in the lane allocation strategy are recalculated based on the real-time parking space occupancy status.

[0168] The preset tolerance range is a pre-set deviation value used to determine whether the entrance traffic flow data matches the predicted traffic peak period. When the deviation between the entrance traffic flow data and the predicted traffic peak period exceeds the preset tolerance range, it means that the prediction is inaccurate and the lane allocation strategy needs to be adjusted. The priority vehicle type and release interval duration in the lane allocation strategy are recalculated based on the real-time parking space occupancy status. Priority vehicle types can include emergency vehicles, public transportation vehicles, etc., and the release interval duration refers to the time interval between the release of adjacent vehicles. For example, the predicted entrance traffic flow during the peak traffic period should be large, but the actual entrance traffic flow is small, and the real-time parking space occupancy status shows that there are many vacancies. At this time, the priority vehicle type can be adjusted to all vehicles, and the release interval duration can be appropriately extended.

[0169] Step S5120: Logically bind the recalculated lane allocation strategy with the signal light phase switching rule to generate a dynamically adjustable composite control instruction set.

[0170] The recalculated lane assignment strategy is logically bound to the signal light phase switching rules, enabling them to work together to form a dynamically adjustable composite control instruction set. This composite control instruction set incorporates comprehensive information about lane assignment and signal light control, enabling more effective traffic flow control at service area entrances. For example, the recalculated lane assignment strategy prioritizing emergency vehicles with a 10-second release interval is bound to the signal light phase switching rule that extends the green light duration to create a composite control instruction set.

[0171] Step S5130: Send a composite control instruction set to the entrance signal light and variable message board through the communication module of the service area control system, and monitor the traffic diversion efficiency after the instruction execution in real time.

[0172] The service area control system's communication module sends a composite control instruction set to the entrance signal lights and variable message signs. The entrance signal lights adjust their phase according to the instruction set, and the variable message signs update their display accordingly. Simultaneously, traffic flow management efficiency after the instruction set is executed is monitored in real time. Traffic flow management efficiency can be measured by metrics such as average waiting time and passing time. For example, after the composite control instruction set is sent to the entrance signal lights and variable message signs, the average waiting time for vehicles passing through the service area entrance is monitored in real time to assess traffic flow management efficiency.

[0173] Step S5140: If the traffic diversion efficiency is lower than the expected threshold, the hovering monitoring platform is triggered to perform a second scan of the road sections around the service area to obtain the updated vehicle density characteristics and regenerate the dynamic traffic control parameters.

[0174] The expected threshold is a pre-set indicator of traffic diversion efficiency. When the traffic diversion efficiency is lower than the expected threshold, it means that the current set of composite control instructions may not achieve the desired effect. At this time, the hovering monitoring platform is triggered to conduct a second scan of the road sections surrounding the service area. The multi-directional sensor array of the hovering monitoring platform scans the road sections surrounding the service area again to obtain updated information such as vehicle density characteristics. Based on the updated information, the dynamic flow control parameters are regenerated, and the traffic strategy at the service area entrance is adjusted again to improve traffic diversion efficiency. For example, if real-time monitoring shows that the average waiting time for vehicles passing through the service area entrance is long, and the traffic diversion efficiency is lower than the expected threshold, the hovering monitoring platform is triggered to conduct a second scan of the road sections surrounding the service area, and the dynamic flow control parameters are regenerated based on the newly acquired vehicle density characteristics, and the lane allocation strategy and signal light phase switching rules are adjusted.

[0175] In one embodiment, in step S500, dynamically controlling the traffic strategy at the service area entrance may specifically include the following steps S5150 to S5210:

[0176] Step S5150: Deploy multi-dimensional sensing equipment at the entrance of the service area to collect vehicle identification data, passenger status data and destination tag data in real time.

[0177] Multi-dimensional sensing equipment, including cameras, sensors, and radio frequency identification (RFID) readers, is deployed at the entrance to the service area. Cameras can identify vehicle features and license plate numbers, sensors can detect vehicle weight, speed, and other information, and RFID readers can read information from RFID tags installed on vehicles. These multi-dimensional sensing devices collect vehicle identification data, passenger status data, and destination tag data in real time. Vehicle identification data uniquely identifies each vehicle entering the service area, passenger status data reveals the vehicle's passenger load, and destination tag data reveals the vehicle's destination. For example, cameras and RFID readers can be installed at the entrance of a service area. When a vehicle enters the service area, the camera recognizes the vehicle's license plate number as identification data, analyzes images of the occupants to obtain passenger status data, and the RFID reader reads the vehicle's RFID tag to obtain destination tag data.

[0178] Step S5160: Filter out emergency rescue vehicles and public transportation vehicles based on vehicle identification data, and correlate and analyze their passenger status data with destination tag data to generate a right of way verification result.

[0179] Based on the collected vehicle identification data, emergency vehicles and public transportation vehicles are screened out. Emergency vehicles usually have specific identification, such as ambulances and fire trucks; public transportation vehicles can be identified by vehicle type and operating information. The passenger status data of the screened emergency vehicles and public transportation vehicles are correlated with the destination tag data for analysis. For example, for an ambulance, its passenger status is analyzed to see if it has a patient, and whether the destination tag data is a hospital. Through correlation analysis, a right-of-way verification result is generated to determine whether these vehicles have the right of way. If an ambulance has a patient and its destination is a hospital, the verification result indicates that it has the right of way.

[0180] Step S5170: Based on the right-of-way verification result, the green light duration and lane guidance direction in the signal light phase switching rule are adjusted to form a hierarchical release strategy.

[0181] Based on the right-of-way verification results, the green light duration and lane guidance in the signal light phase switching rules are adjusted. For vehicles with right-of-way, the green light duration is extended to ensure they can quickly pass through the service area entrance. At the same time, lane guidance is adjusted to provide dedicated lanes for priority vehicles. A tiered release strategy is implemented, classifying vehicles as priority and non-priority vehicles, with different release rules applied to different vehicle levels. For example, for ambulances with right-of-way, the green light duration of the signal light is extended from 20 seconds to 30 seconds, and a dedicated lane is set up to guide them quickly into the service area.

[0182] Step S5180: For non-priority vehicles, diversion suggestion information is generated based on the matching degree between their destination tag data and the remaining capacity of adjacent service areas and is dynamically displayed through a variable information board.

[0183] For non-priority vehicles, the matching degree between their destination tag data and the remaining capacity of adjacent service areas is analyzed. The remaining capacity of adjacent service areas can be obtained in real time through the service area control system. If the remaining capacity of a service area near the destination of a non-priority vehicle is small, while the remaining capacity of another adjacent service area is large, a diversion suggestion information is generated to guide the vehicle to the service area with larger remaining capacity. These diversion suggestion information are dynamically displayed on the variable information board to provide reasonable guidance for non-priority vehicles. For example, the destination tag data of a non-priority vehicle shows that it is going to service area A, but the remaining capacity of service area A is small, while the remaining capacity of the adjacent service area B is large. The variable information board displays the diversion suggestion information "It is recommended to go to service area B, there are sufficient parking spaces remaining."

[0184] Step S5190: Real-time statistics are collected on the vehicle response ratio to the diversion suggestion information. If the response ratio is lower than the diversion effectiveness threshold, the signal light waiting period of non-priority vehicles is shortened and the push frequency of the diversion suggestion information is increased.

[0185] Real-time statistics are collected on the vehicle response ratio to the diversion suggestion information, that is, the ratio of the number of vehicles traveling according to the diversion suggestion information to the number of vehicles receiving the diversion suggestion information. The diversion effectiveness threshold is a pre-set ratio value. When the vehicle response ratio is lower than the diversion effectiveness threshold, it means that the diversion suggestion information is not effective. At this time, the signal light waiting period of non-priority vehicles is shortened so that non-priority vehicles can pass through the service area entrance faster; at the same time, the push frequency of diversion suggestion information is increased to improve the dissemination effect of the information. For example, real-time statistics show that the vehicle response ratio to the diversion suggestion information is 30%, which is lower than the diversion effectiveness threshold of 50%. The signal light waiting period of non-priority vehicles is shortened from 30 seconds to 20 seconds, and the push frequency of diversion suggestion information is increased from once every 5 minutes to once every 3 minutes.

[0186] Step S5200: When it is detected that the traffic volume at the entrance of the service area drops to a safe diversion range, the signal light phase switching rules are gradually restored to the baseline state, and the diversion suggestion information in the variable information board is turned off.

[0187] When the service area control system detects that the traffic volume at the service area entrance has dropped to a safe diversion range, it means that the traffic pressure has been reduced and no special traffic strategy is needed. At this time, the signal light phase switching rules are gradually restored to the baseline state, which refers to the signal light phase switching rules under normal circumstances. At the same time, the diversion suggestion information on the variable information board is turned off, and the traffic at the service area entrance returns to normal. For example, when it is detected that the traffic volume at the service area entrance has dropped from 500 vehicles per hour during peak hours to 100 vehicles per hour, which is within the safe diversion range, the signal light phase switching rules are gradually restored to the normal green light time and switching sequence, and the diversion suggestion information on the variable information board is turned off.

[0188] Step S5210: Feedback the execution effect of the graded release strategy to the command center through the satellite communication unit, triggering the hovering monitoring platform to update the multi-level warning signal generation logic for subsequent abnormal events.

[0189] The effectiveness of the tiered release strategy is fed back to the command center via satellite communications. This effectiveness can include metrics such as traffic flow management efficiency and vehicle response rates. Upon receiving this feedback, the command center evaluates the effectiveness of the tiered release strategy. Based on the evaluation results, the hovering monitoring platform is triggered to update the logic for generating multi-level warning signals for subsequent abnormal events. For example, if the tiered release strategy performs well and traffic flow management efficiency is significantly improved, the hovering monitoring platform can increase its focus on priority vehicles in subsequent abnormal events and optimize the rules for generating multi-level warning signals, ensuring more efficient traffic management.

[0190] It is understandable that the various algorithms involved in the above-mentioned introductions of the embodiments of the present invention, such as edge detection algorithms, CRC algorithms, conflict resolution algorithms, etc., can all be learned from the relevant content in the prior art. In order to save space, they will not be expanded too much in the embodiments of the present invention. In addition, when implementing the scheme of the present invention, those skilled in the art can supplement the details according to the common knowledge in this field. For example, according to the common knowledge in this field, normalization can be used to eliminate dimensional conflicts before feature fusion, interpolation can be used to eliminate dimensional differences, and historical data, experience or business scenario requirements can be combined to reasonably set thresholds, and models can be trained based on general model training methods, etc. The present invention will no longer provide redundant introductions to overly detailed implementation processes.

[0191] See also Figure 2 , Figure 2This is a schematic diagram of the structure of a monitoring system provided in an embodiment of the present invention. The monitoring system can be a backend server system of a traffic management department, used to analyze monitoring data uploaded by data acquisition equipment and issue control instructions. Specifically, the monitoring system includes at least a processor 101, a communication interface 102, and a memory 103. The processor 101, communication interface 102, and memory 103 can be connected via a bus or other means. The processor 101 (also known as the Central Processing Unit (CPU)) is the computing and control core of the monitoring system, capable of parsing various instructions within the monitoring system and processing various types of data within the monitoring system. The communication interface 102 can optionally include a standard wired interface or a wireless interface (such as Wi-Fi, a mobile communication interface, etc.), and can be used to send and receive data under the control of the processor 101. The communication interface 102 can also be used for data transmission and interaction within the monitoring system. The memory 103 is a storage device in the monitoring system, used to store programs and data. It is understood that the memory 103 here can include both the built-in memory of the monitoring system and the extended memory supported by the monitoring system. Memory 103 provides storage space for storing the monitoring system's operating system, which is not limited in the present invention. In one embodiment, processor 101 executes the drone cruise warning method for high-speed traffic monitoring provided in the above embodiments of the present invention by running the computer program in memory 103.

Claims

1. A drone cruise warning method for high-speed traffic monitoring, characterized in that: include: The system collects dynamic monitoring data of the target road section in real time through multiple hovering monitoring platforms deployed along the highway, wherein each of the hovering monitoring platforms includes an energy supply module and a signal relay module; Identifying abnormal events existing in a target road section based on the dynamic monitoring data, and generating multi-level warning signals according to the event type of the abnormal event; Transmitting the multi-level warning signal to a command center, triggering the command center to generate a drone dispatch instruction matching the event type, the drone dispatch instruction including the target drone's identifier and path planning parameters; Controlling the target UAV to move to a target location corresponding to the abnormal event based on the path planning parameters, and executing a warning operation corresponding to the event type through the external speaker module of the target UAV; Calculate the diffusion rate of the abnormal event and the impact boundary change trend based on the event processing status fed back by the command center and the real-time status data of the target UAV; If the diffusion rate exceeds the capture capability corresponding to the current monitoring coverage radius of the hovering monitoring platform, a collaborative monitoring protocol of adjacent hovering monitoring platforms is activated, wherein the collaborative monitoring protocol includes data synchronization frequency adjustment rules and sensor array scanning angle complementation strategy; The dynamic monitoring data of multiple hovering monitoring platforms are temporally and spatially aligned and fused through the collaborative monitoring protocol to generate an extended monitoring data set, wherein the extended monitoring data set includes cross-platform vehicle trajectory continuity characteristics and environmental parameter superposition effect data; recalculating the spatial distribution gradient of the vehicle density feature based on the extended monitoring data set, and identifying newly added congestion nodes or accident-derived areas in the impact boundary change trend; Based on the location coordinates of the newly added congested node or accident-derived area, adjusting the scanning focus direction of the multi-directional sensor array of the hovering monitoring platform and increasing the image acquisition frame rate of the newly added congested node; The adjusted scanning focus direction and image acquisition frame rate parameters are encapsulated as a monitoring range update instruction and sent to all hovering monitoring platforms participating in the collaborative monitoring protocol; receiving, in real time, feedback on the status of each platform executing the monitoring range update command via a satellite communication unit; and if it is detected that some platforms have not completed parameter update, activating a backup hovering monitoring platform to replace the platform that has not responded to the command and reallocating its monitoring coverage area; Based on the vehicle trajectory continuity characteristics in the extended monitoring data set, predict the probability distribution of target direction vehicles arriving at the service area within a preset time period and the traffic peak period; Generate dynamic traffic control parameters for the service area entrance based on the probability distribution and traffic peak period, the dynamic traffic control parameters including entrance lane allocation strategy, signal light phase switching rules and variable information board content update cycle; Acquire real-time entrance traffic flow data and parking space occupancy status through the service area control system, and perform matching analysis with the dynamic traffic control parameters; If the deviation between the entrance traffic flow data and the predicted traffic peak period exceeds a preset tolerance range, the priority vehicle type and release interval duration in the lane allocation strategy are recalculated based on the real-time parking space occupancy status; Logically bind the recalculated lane allocation strategy with the signal light phase switching rules to generate a dynamically adjustable composite control instruction set; Sending the composite control instruction set to the entrance signal lights and variable message boards through the communication module of the service area control system, and monitoring the traffic flow diversion efficiency after the execution of the instructions in real time; If the traffic diversion efficiency is lower than the expected threshold, the hovering monitoring platform is triggered to perform a second scan of the road sections around the service area to obtain updated vehicle density characteristics and regenerate the dynamic traffic control parameters.

2. The method according to claim 1, characterized in that The hovering monitoring platform is equipped with a wide-area monitoring module within a preset height range. The wide-area monitoring module includes a multi-directional sensor array and a satellite communication unit. The multiple hovering monitoring platforms deployed along the highway collect dynamic monitoring data of the target road section in real time, including: Performing omnidirectional scanning of a target road section through the multi-directional sensor array to obtain real-time image data and environmental parameter data of the target road section; Inputting the real-time image data into a preset event recognition model, outputting dynamic trajectory features of vehicle behavior in the target road section, and comparing the dynamic trajectory features with a preset safety threshold; If the dynamic trajectory feature exceeds the safety threshold, a preliminary warning signal is sent to the command center via the satellite communication unit, and an abnormal timestamp is marked in the local storage module of the hovering monitoring platform; The dynamic monitoring data is generated according to the correlation between the environmental parameter data and the abnormal timestamp, and the dynamic monitoring data includes the time synchronization tag of the real-time image data and the spatial distribution information of the environmental parameter data.

3. The method according to claim 2, characterized in that The identifying abnormal events existing in the target road section based on the dynamic monitoring data includes: Extracting vehicle density features, vehicle speed deviation features, and lane occupancy features from the dynamic monitoring data; Matching the vehicle density feature with the peak density in a historical traffic database to determine a first congestion index of the target road section; calculating a speed difference gradient between adjacent vehicles based on the vehicle speed deviation characteristic, and generating a second congestion index if the speed difference gradient exceeds a preset gradient threshold; Analyzing the occupancy status of the emergency lane based on the lane occupancy characteristics, and generating a third congestion index if the occupancy time of the emergency lane exceeds a preset response time limit; The first congestion index, the second congestion index, and the third congestion index are input into an event classification model, and an event type of the abnormal event is output, where the event type includes at least one of an accident type, a congestion type, and an obstacle type.

4. The method according to claim 3, characterized in that Generating a multi-level warning signal according to the event type of the abnormal event includes: Determine a target coverage range of the warning signal based on the event type, wherein the target coverage range includes the lane where the abnormal event occurred, adjacent lanes, and associated road sections within a preset distance in the target direction; Generate dynamic warning radiation parameters according to the correlation between the vehicle density characteristics and the target coverage range, wherein the dynamic warning radiation parameters include signal propagation direction, signal strength attenuation gradient, and signal priority weight; Sending a pre-carrier detection signal to vehicles within the target coverage area through the signal relay module of the hovering monitoring platform, and receiving a carrier response signal fed back by the vehicle; Adjusting the signal strength attenuation gradient in the dynamic warning radiation parameter according to the signal quality parameter of the carrier response signal, and generating a directional enhancement signal based on the adjusted signal strength attenuation gradient; Encapsulating the event type, target coverage, and directional enhancement signal into a multi-level warning signal, and temporally and spatially correlating the multi-level warning signal with the geographic location tag of the hovering monitoring platform to generate an encrypted data packet; sending the encrypted data packet to the command center via the satellite communication unit, and monitoring the transmission integrity of the encrypted data packet in real time during the sending process; If it is detected that the transmission integrity is lower than the preset integrity threshold, the local storage module of the hovering monitoring platform is triggered to temporarily store the multi-level warning signal, and retransmit the encrypted data packet through the backup communication link until the command center returns a reception confirmation signal.

5. The method according to claim 4, characterized in that The triggering of the command center to generate a drone dispatch instruction matching the event type includes: Based on the event type and target coverage in the multi-level warning signal, a set of candidate drones with corresponding functional modules are screened from the drone resource pool; Calculating a scheduling priority score for each UAV in the candidate UAV set based on its endurance parameter, current mission load status, and relative distance to the target location; Sorting the set of candidate drones according to the scheduling priority scores, and selecting the drone with the highest score as the target drone; Extracting the signal propagation direction and signal strength attenuation gradient from the dynamic warning radiation parameters to generate initial path planning parameters for the target UAV; Based on the real-time environmental parameter data of the target location, the initial path planning parameters are dynamically modified to obtain path planning parameters including obstacle avoidance path nodes, flight altitude gradient, and return charging strategy; Encapsulate the target drone's identification, path planning parameters, and event processing requirements of the multi-level warning signal into instructions to generate the drone dispatch instruction; The energy supply module of the hovering monitoring platform provides a pre-charging interface for the target UAV, and verifies its compatibility with the path planning parameters before the target UAV takes off.

6. The method according to claim 5, characterized in that The controlling the target UAV to move to the target location corresponding to the abnormal event based on the path planning parameters includes: After the target UAV takes off, a navigation calibration signal is continuously sent to the target UAV through the signal relay module of the hovering monitoring platform, wherein the navigation calibration signal includes a target position offset and an environmental interference compensation parameter updated in real time; Adjusting the flight altitude gradient of the target UAV based on the environmental interference compensation parameter, and collecting attitude data and energy consumption data of the target UAV in real time; Comparing the posture data with the standard posture range in the path planning parameters, and if it is detected that the posture deviation exceeds the fault tolerance threshold, generating a dynamic obstacle avoidance path according to the obstacle avoidance path node; Calculating the remaining flight time of the target UAV according to the energy consumption data, and adjusting the charging priority in the return charging strategy based on the remaining flight time; When the target UAV approaches the target location, the real-time distance data to the target location is obtained through the positioning module carried by the target UAV; If the real-time distance data is less than a preset trigger threshold, the external speaker module of the target drone is activated to enter a warning ready state, and the image acquisition module is activated simultaneously; The real-time status data of the target UAV is transmitted back to the command center through the signal relay module, triggering the command center to update the path planning parameters in the UAV scheduling instruction.

7. The method according to claim 6, characterized in that The performing of the warning operation corresponding to the event type by the external speaker module of the target UAV includes: Calling a preset warning mode library of the external speaker module according to the event type, the warning mode library includes an audio and light warning combination, a voice broadcast template and a laser positioning mode; If the event type is an accident type, the directional sound wave pulse and strobe light sequence in the sound and light warning combination are selected, and the image acquisition module is simultaneously started to record a panoramic video stream of the accident scene; If the event type is congestion, a multilingual alternative route prompt in the voice broadcast template is selected, and a deceleration control instruction is sent to the vehicle terminal of the adjacent vehicle through the communication module of the target UAV; If the event type is an obstacle type, activating the laser positioning mode to project an obstacle boundary mark and generate three-dimensional contour data of the obstacle; Transmitting the panoramic video stream, deceleration control instructions, and three-dimensional contour data to the command center in real time, triggering the command center to update the event processing status of the multi-level warning signal according to the panoramic video stream; When the event processing status changes to completed, controlling the target UAV to return to the nearest hovering monitoring platform based on the adjusted return-to-home charging strategy; During the return process, the remaining flight time of the target UAV is continuously monitored. If the remaining flight time is lower than the safety threshold, the emergency charging protocol is activated and mid-flight charging is performed through the hovering monitoring platform passing by.

8. A monitoring system, characterized in that: include: a memory storing a computer program; A processor, configured to load the computer program to implement the drone cruise warning method for high-speed traffic monitoring as described in any one of claims 1 to 7.

Images