Artificial intelligence-based network optimization method and background system

Abstract

The present disclosure provides an artificial intelligence-based network optimization method, a background system, equipment and a medium, the method comprising: S1, according to the three-dimensional map of the target area, the patrol airspace is divided into low-altitude layer, medium-altitude layer and high-altitude layer according to height. Layered deployment is performed on multiple patrol unmanned aerial vehicles, and network configuration and channel planning are performed. S2, the ideal signal strength is calculated, the ideal signal strength is compared with the actual signal strength, and it is judged whether the actual signal meets the standard. S3, if it is determined that the actual signal does not meet the standard, it is determined that there is an interference source, and the positioning or motion trail of the interference source and the interference processing scheme are reported. The present disclosure can improve the inspection efficiency, complete the whole process through spectrum acquisition, artificial intelligence identification, triangular positioning and dynamic tracking, and the background system only identifies, positions, generates a scheme and reports, does not intervene in the bottom layer execution, the system architecture is clear, and the landing performance is strong.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence-based network optimization technology, and in particular to an artificial intelligence-based network optimization method, a backend system, an electronic device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] With the rapid development of the low-altitude economy, the application of equipment such as drones and low-altitude aircraft is becoming increasingly widespread. Their reliable operation is highly dependent on the stability of low-altitude wireless communication links (such as flight control signals and data feedback signals). In existing technologies, low-altitude wireless environmental perception often adopts a single drone and fixed-gain antenna switching mode. For example, a single drone carries multiple antenna units with different gains to scan signals along a fixed flight path, and switches antennas or issues alarms after comparing the signal strength with the ideal signal strength.

[0003] However, a single drone can only perform single-point detection along a fixed flight path, failing to simultaneously cover wireless environments at different altitudes and in different areas. For large low-altitude areas (such as urban core areas and logistics route clusters), inspection time is more than three times the ideal value, and blind spots are easily created. Existing technologies can only detect the result of poor signal, but cannot identify the root cause of signal degradation (such as electromagnetic interference, physical obstruction, and temporary interference sources), let alone locate mobile interference sources (such as ground electromagnetic interference vehicles and temporary WiFi amplifiers), resulting in low troubleshooting efficiency. Some existing technologies only combine distance and meteorological data to calculate ideal signal strength, without considering the signal loss caused by obstacles (such as tall buildings, trees, and bridges) in the low-altitude three-dimensional environment. The deviation between the ideal signal and the actual scene can reach 10-15 dBm, leading to a high false positive rate. In addition, existing technologies can only detect signal degradation in real time and cannot predict signal changes in future time periods (such as 6-24 hours) (such as signal attenuation caused by rainfall and frequency congestion caused by peak traffic), which can easily lead to sudden communication interruptions during drone flight.

[0004] In other words, existing low-altitude wireless environment sensing methods suffer from problems such as low inspection efficiency, difficulty in locating interference sources, lack of signal prediction, poor accuracy of ideal signals, and limited antenna switching strategies. Summary of the Invention

[0005] This disclosure provides a network optimization method based on artificial intelligence, a backend system, an electronic device, a computer-readable storage medium, and a computer program product.

[0006] Firstly, this disclosure provides an artificial intelligence-based network optimization method applied to a backend system, the method comprising: S1, based on the 3D map of the target area, divides the inspection airspace into low-altitude, mid-altitude, and high-altitude layers according to altitude. Multiple inspection drones are then deployed in layers, with network configuration and flight path planning.

[0007] S2 controls multiple inspection drones to fly along their respective flight paths, receiving meteorological data from the high-altitude drone group and lidar point cloud data and actual signal strength from the low-altitude drone group. It calculates the ideal signal strength based on the lidar point cloud data and meteorological data. The ideal signal strength is then compared with the actual signal strength to determine if the actual signal meets the standard.

[0008] S3, if the actual signal is determined to be substandard, and an interference source is identified, then: Control the mid-altitude UAV group to collect and transmit spectrum data of the area where the actual signal is substandard. Identify the type of interference source based on the spectrum data. Determine the interference handling plan based on the type of interference source. Control the low-altitude UAV group to locate the interference source and transmit its location coordinates. If the interference source is dynamic, control the low-altitude UAV group to track the interference source and transmit its tracking trajectory to obtain the movement trajectory of the interference source. Report the location or movement trajectory of the interference source and the interference handling plan.

[0009] In some embodiments, S1, the hierarchical deployment of multiple inspection drones includes: deploying a low-altitude drone group in the low-altitude layer, a medium-altitude drone group in the medium-altitude layer, and a high-altitude drone group in the high-altitude layer. Each drone in the low-altitude drone group is equipped with a high-precision spectrum analyzer, a lidar, and a low-gain antenna unit. Each drone in the medium-altitude drone group is equipped with a spectrum data acquisition module, a medium-gain antenna unit, and a dual-mode communication module. Each drone in the high-altitude drone group is equipped with a long-range signal receiver, a multi-parameter weather sensor, and a high-gain antenna unit. The network configuration of the multiple inspection drones includes: using the high-altitude drone group as a core node, enabling it to receive data from the medium-altitude drone group and transmit it back to the backend system; and using the medium-altitude drone group as a relay node, enabling data collected by the low-altitude drone group to be forwarded to the high-altitude drone group via the medium-altitude drone group.

[0010] In some embodiments, S2, the method for obtaining the ideal signal strength includes: S21, calculates occlusion loss based on lidar point cloud data.

[0011] S22, calculate meteorological losses based on meteorological data.

[0012] S23. Calculate the ideal signal strength based on the base station's transmit power, obstruction loss, weather loss, base station antenna gain, and the antenna gain of each UAV in the low-altitude UAV group.

[0013] In some embodiments, S21 includes: denoising the lidar point cloud data to obtain denoised point cloud data, and using an ICP algorithm to obtain a three-dimensional mesh model based on the denoised point cloud data. Obstacles in the three-dimensional mesh model are classified according to an artificial intelligence recognition model, and a preset dielectric constant is matched to each obstacle according to its type. The thickness of each obstacle was calculated based on the 3D mesh model. According to the formula Calculate the shading loss ,in, It is the wavelength of the signal emitted by the base station.

[0014] In some embodiments, S22 includes: obtaining a meteorological loss coefficient based on meteorological data and a meteorological model, and multiplying the meteorological loss coefficient by the actual transmission distance of the signal to obtain the meteorological loss. The actual transmission distance of the signal is the distance between the base station and the target drone, and the target drone is the drone to which the actual signal strength belongs.

[0015] In some embodiments, S23 includes: S231, according to formula Calculate free space loss .

[0016] S232, according to formula Calculate the ideal signal strength ,in, It is the base station's transmission power. It is free space loss. It is weather loss. It's due to shading loss. It is the base station antenna gain. It is the antenna gain of each UAV in the low-altitude UAV group.

[0017] In some embodiments, in S3, the mid-altitude UAV group is controlled to collect and transmit the spectrum data of the interference source. Identifying the interference source type based on the spectrum data includes: controlling the mid-altitude UAV group to collect and transmit the spectrum data of the signal-deficient region; extracting spectral features from the transmitted spectrum data using a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) model, and analyzing the signal time series changes to determine the interference type. Controlling the low-altitude UAV group to locate the interference source and transmit its location coordinates includes: scheduling three UAVs in the low-altitude UAV group to construct a triangulation network; calculating and transmitting the three-dimensional coordinates of the interference source based on the angle of arrival and time difference of the interference signal transmitted by each UAV in the triangulation network.

[0018] In some embodiments, the method further includes: acquiring flight data and pedestrian flow data of the low-altitude unmanned aerial vehicle (UAV) group. The flight data, pedestrian flow data, and meteorological data of the low-altitude UAV group are input into a prediction model to obtain a predicted value of the signal strength within a preset future time period. The prediction model is based on historical data of the low-altitude UAV group, historical pedestrian flow data, historical meteorological data, and a long short-term memory network. If the predicted value of the signal strength in the first area within the preset future time period is lower than a preset threshold, then A is applied, using airway length, obstruction loss, and predicted signal strength as cost factors. The algorithm replans the UAV flight paths in the first region, obtaining an optimized flight path scheme, and then distributes the optimized flight path scheme to the relevant UAVs in the low-altitude UAV group, controlling the relevant UAVs to switch flight paths. The first region is a space area in the low-altitude layer.

[0019] In some embodiments, the method further includes: acquiring and recording the signal compliance rate and the coordinates of the weakest signal area in real time, and generating a signal quality report; acquiring and recording the type, quantity, location coordinates, and processing results of interference sources, and generating an interference source report; acquiring interference source processing reports and signal prediction reports for a preset future time period in real time, and generating a prediction report; generating and displaying a 3D signal heatmap, interference source trajectory animation, and signal prediction curve, and configuring corresponding data viewing permissions for different roles.

[0020] Secondly, this disclosure provides a backend system, including: The planning module is configured to divide the inspection airspace into low-altitude, mid-altitude, and high-altitude layers based on a 3D map of the target area. It also enables layered deployment of multiple inspection drones, including network configuration and flight path planning.

[0021] The signal receiving module is configured to control each of the multiple inspection drones to fly along its respective flight path, receiving meteorological data from the high-altitude drone group and lidar point cloud data and actual signal strength from the low-altitude drone group. It calculates the ideal signal strength based on the lidar point cloud data and meteorological data. The ideal signal strength is then compared with the actual signal strength to determine whether the actual signal meets the standard.

[0022] The interference source processing module is configured as follows: If the actual signal is determined to be substandard, indicating the presence of an interference source, then: the mid-altitude UAV group is controlled to collect and transmit spectrum data of the area where the actual signal is substandard. The type of interference source is identified based on the spectrum data. An interference handling scheme is determined based on the interference source type. The low-altitude UAV group is controlled to locate the interference source and transmit its coordinates. If the interference source is dynamic, the low-altitude UAV group is controlled to track the interference source and transmit its tracking trajectory to obtain the movement trajectory of the interference source. The location or movement trajectory of the interference source and the interference handling scheme are then reported.

[0023] Thirdly, this disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores one or more computer programs executable by the at least one processor, the one or more computer programs being executed by the at least one processor to enable the at least one processor to perform the aforementioned artificial intelligence-based network optimization method.

[0024] Fourthly, this disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the aforementioned artificial intelligence-based network optimization method.

[0025] Fifthly, this disclosure provides a computer program product that includes computer-readable code or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device executes the above-described artificial intelligence-based network optimization method.

[0026] The embodiments provided in this disclosure enable large-scale layered coverage, with three layers of drones operating collaboratively at different altitudes, significantly improving inspection efficiency. The inspection time for a 100-square-kilometer area is reduced from four hours to 1.2 hours. Simultaneously, it achieves a qualitative leap from simply knowing the signal is poor to understanding why it's poor, who is interfering, where the interference is occurring, and how to handle it. The entire process is completed through spectrum acquisition, artificial intelligence identification, triangulation, and dynamic tracking. Furthermore, the backend system only performs identification, location, solution generation, and reporting, without interfering with the underlying execution. The system architecture is clear and highly implementable.

[0027] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0028] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the embodiments of the present disclosure to explain the disclosure and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed description of exemplary embodiments with reference to the accompanying drawings, in which: Figure 1 A flowchart illustrating an artificial intelligence-based network optimization method provided in this embodiment of the disclosure; Figure 2 A flowchart illustrating another AI-based network optimization method provided in this disclosure embodiment; Figure 3A flowchart illustrating yet another AI-based network optimization method provided in this disclosure embodiment; Figure 4 A block diagram of a backend system provided in an embodiment of this disclosure; Figure 5 This is a block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0029] To enable those skilled in the art to better understand the technical solutions of this disclosure, exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments of this disclosure to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0030] Where there is no conflict, the various embodiments of this disclosure and the features thereof in the embodiments may be combined with each other.

[0031] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0032] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Words such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0033] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined herein.

[0034] Embodiments of this disclosure provide an artificial intelligence-based network optimization method applied to a backend system, such as... Figure 1 As shown, the method includes steps S1 to S3.

[0035] Step S1: Based on the 3D map of the target area, divide the inspection airspace into low-altitude, mid-altitude, and high-altitude layers according to altitude. Deploy multiple inspection drones in layers, configure their network, and plan their flight paths.

[0036] In some embodiments, step S1, the hierarchical deployment of multiple inspection drones, includes: deploying a low-altitude drone group in the low-altitude layer, a medium-altitude drone group in the medium-altitude layer, and a high-altitude drone group in the high-altitude layer. Each drone in the low-altitude drone group is equipped with a high-precision spectrum analyzer, a lidar, and a low-gain antenna unit. Each drone in the medium-altitude drone group is equipped with a spectrum data acquisition module, a medium-gain antenna unit, and a dual-mode communication module. Each drone in the high-altitude drone group is equipped with a long-range signal receiver, a multi-parameter weather sensor, and a high-gain antenna unit. The network configuration of the multiple inspection drones includes: using the high-altitude drone group as a core node, so that the high-altitude drone group receives data from the medium-altitude drone group and then transmits it back to the backend system; and using the medium-altitude drone group as a relay node, so that data collected by the low-altitude drone group is forwarded to the high-altitude drone group via the medium-altitude drone group.

[0037] For example, the target inspection area can be divided into three levels according to altitude: low altitude (50-200m), mid altitude (200-500m) and high altitude (500-800m).

[0038] The drones deployed in the low-altitude layer are responsible for high-precision signal acquisition in core areas (such as logistics routes and key inspection areas). The equipment they carry includes: a high-precision spectrum analyzer (bandwidth 100kHz-26.5GHz, signal strength measurement accuracy ±0.5dBm), a lidar (point cloud density 300 points / second, ranging accuracy ±2cm), and a low-gain antenna unit (gain 1dBi, compatible with 2.4G / 5.8G / 4G frequency bands, simulating the receiving capabilities of ordinary drones).

[0039] The drones deployed in the mid-altitude layer are responsible for the initial identification of interference sources and regional signal coverage. The equipment they carry includes: a spectrum data acquisition module; a medium-gain antenna unit (gain 3dBi, supports adaptive bandwidth adjustment, bandwidth range 1-20MHz); and a dual-mode communication module (millimeter wave 28GHz band, latency ≤10ms; LoRa 868MHz band, spreading factor 12, used for long-distance backup).

[0040] The UAVs deployed in the high-altitude layer are responsible for large-scale signal monitoring and meteorological data collection. The equipment they carry includes: a long-range signal receiver (receiving sensitivity -97dBm); a multi-parameter meteorological sensor (measuring temperature -40~85℃, humidity 0~100%RH, rainfall 0~50mm / h); and a high-gain antenna unit (gain 9dBi, supports beamforming, coverage angle 360°).

[0041] For example, communication is established through a tree-like network structure of high-altitude UAVs, medium-altitude UAVs, and low-altitude UAVs. The high-altitude UAVs serve as the network core, communicating with the ground monitoring backend (i.e., the backend system) via 4G / 5G public networks. The medium-altitude UAVs act as relay nodes, forwarding data collected by the low-altitude UAVs. The medium-altitude UAV group sends two types of data to the high-altitude UAVs: data from the low-altitude UAV group and data collected by the medium-altitude UAV group itself.

[0042] For example, waterway planning can be based on a 3D map of the target area (such as Gaode Maps). Figure 3 (API acquisition) is used to plan collaborative flight paths for drones at various levels. In the low-altitude layer, zigzag flight paths are set up along core routes (such as fixed routes for logistics drones) with a spacing of 50m to ensure no detection blind spots. In the mid-altitude layer, a ring-shaped flight path is set up 100m outside the low-altitude flight path to cover the signal radiation range of the low-altitude layer. In the high-altitude layer, a grid-shaped flight path is set up along the boundary of the target area with a grid spacing of 500m to achieve wide-area signal monitoring.

[0043] Understandably, this configuration ensures a clear division of labor and distinct responsibilities among the various drone levels: low-altitude drones handle high-precision signal acquisition and environmental scanning, mid-altitude drones handle spectrum data acquisition and data forwarding, high-altitude drones handle data aggregation and meteorological data collection, and the backend handles all intelligent computing. This distributed collaborative architecture significantly enhances the dimensionality and comprehensiveness of data acquisition, while the tree-like network ensures stable data transmission. The measured transmission rate reaches 15 megabits per second, with a packet loss rate of only 0.05%.

[0044] Step S2: Control each of the multiple inspection drones to fly along its respective flight path, receiving meteorological data from the high-altitude drone group and lidar point cloud data and actual signal strength from the low-altitude drone group. Calculate the ideal signal strength based on the lidar point cloud data and meteorological data. Compare the ideal signal strength with the actual signal strength to determine whether the actual signal meets the standard.

[0045] For example, a low-altitude UAV activates its low-gain antenna unit and flies along a planned flight path, collecting wireless signal data once per second, including signal strength (dBm), frequency (MHz), modulation method (e.g., OFDM, FSK), cell number, and signal-to-noise ratio (SNR, dB). A mid-altitude UAV activates its medium-gain antenna unit to simultaneously collect signal data from the periphery of the low-altitude flight path and forwards the collected signals to a high-altitude UAV in real time. A high-altitude UAV activates its high-gain antenna unit to collect signal data over a wide area and simultaneously collects real-time meteorological data (temperature, humidity, rainfall) via meteorological sensors, and collects three-dimensional environmental data (building height, tree distribution) via lidar.

[0046] For example, a method for determining whether the actual signal meets the standard may include: setting a signal strength threshold (e.g., T1); if the difference between the ideal signal strength and the actual signal strength is less than the preset signal strength threshold (T1), then the actual signal meets the standard; otherwise, it does not meet the standard.

[0047] In some embodiments, such as Figure 2 As shown, in step S2, the method for obtaining the ideal signal strength includes steps S21 to S23.

[0048] Step S21: Calculate the occlusion loss based on the lidar point cloud data.

[0049] In some embodiments, step S21 includes: denoising the lidar point cloud data to obtain denoised point cloud data, and using the ICP algorithm to obtain a three-dimensional mesh model based on the denoised point cloud data. Obstacles in the three-dimensional mesh model are classified according to an artificial intelligence recognition model, and a preset dielectric constant is matched to each obstacle according to its type. The thickness d of each obstacle is calculated based on the three-dimensional mesh model, and the shading loss is calculated according to equation (1). .

[0050] (1)

[0051] In equation (1), λ is the wavelength of the signal emitted by the base station or the signal emitted by the flight controller.

[0052] Understandably, base stations emit signals at different frequencies. Therefore, when a low-altitude drone collects data, it will receive signals from different frequency bands, such as 2.4GHz flight control signals, 5.8GHz data transmission signals, or 4G LTE signals. When calculating obstruction loss, the wavelength λ in the formula is calculated based on the frequency of the signal in that specific frequency band. For example, if analyzing a 2.4GHz signal, the wavelength is 0.125 meters; if switching to 5.8GHz, the wavelength becomes approximately 0.052 meters. Signals of different frequencies experience different losses when passing through the same obstacle, so the wavelength of the corresponding signal must be used for calculation.

[0053] For example, a low-altitude UAV transmits point cloud data (format: PCD) collected by lidar to a ground-based backend, where PCL (Point Cloud Library) is used for preprocessing. Voxel filtering: Set the voxel size to 5cm to remove redundant point clouds and reduce the amount of data; ICP registration: Aligning point clouds from multiple frames to generate a 3D mesh model of the target region (resolution 0.1m). Obstacle Classification: Based on a CNN (Convolutional Neural Network) recognition model, this system classifies obstacles in a 3D model (concrete skyscrapers, wooden trees, metal bridges) and matches the dielectric constant of each type of obstacle (concrete). =6.5, wood =1.5, metal →∞).

[0054] Understandably, introducing 3D environment modeling and obstacle classification into low-altitude wireless environment perception significantly reduces the deviation between the ideal signal and the actual scene from 10 to 15 milliwatt decibels, and significantly lowers the signal misjudgment rate.

[0055] Step S22: Calculate meteorological losses based on meteorological data.

[0056] In some embodiments, step S22 includes: obtaining a meteorological loss coefficient based on meteorological data and a meteorological model, and multiplying the meteorological loss coefficient by the actual transmission distance of the signal to obtain the meteorological loss. The actual transmission distance of the signal is the distance between the base station and the target drone, and the target drone is the drone to which the actual signal strength belongs.

[0057] For example, the straight-line distance D (in km) between the UAV and the currently connected base station can be determined based on the UAV's location information (GPS positioning accuracy ±0.5m). The meteorological model can refer to the ITU-R P.676 model; meteorological losses can be calculated using the ITU-R P.676 model based on meteorological data. (For example, when the rainfall is 5 mm / h, the meteorological loss in the 2.4 GHz band is 2 dB / km).

[0058] Understandably, the ITU-R P.676 model, officially titled "Attenuation in Atmospheric Gases," is an international standard recommendation published by the Radiocommunication Sector of the International Telecommunication Union (ITU). It provides a widely accepted mathematical model in the global communications industry for calculating signal attenuation caused by the absorption of radio waves by oxygen and water vapor in the Earth's atmosphere as they pass through it. This model covers an extremely wide frequency range from 1 GHz to 1000 GHz. When calculating meteorological attenuation, real-time data such as temperature, humidity, and air pressure measured by a drone are input into this model to obtain a signal attenuation value (in dB / km) for a specific frequency band and under specific meteorological conditions.

[0059] Understandably, for the main meteorological factors causing signal attenuation (such as large-scale rainfall and wide-ranging humidity changes), the changes are usually not drastic within a vertical height difference of several hundred meters. The vertical thickness of a rain zone is often much greater than 500 meters; if rain is detected at high altitudes, it is highly likely that it is also raining at low altitudes, and the attenuation coefficient can be considered basically the same. The low-altitude region of 50-200 meters is precisely where the ground heat island effect, vehicle exhaust, and local dust are most complex. The temperature and humidity measured at this point are highly localized and cannot represent the average loss of the signal when traversing a long path. The values ​​measured at high altitudes are more representative of the macroscopic background state of the atmosphere in that area, and are more stable and reliable for calculating path loss. Therefore, in this disclosure, it is assumed that the meteorological loss in the high-altitude layer is consistent with that in the low-altitude layer.

[0060] This disclosure multiplies the meteorological loss coefficient by the actual transmission distance, making the quantification of meteorological impact more accurate, and is especially suitable for signal assessment under adverse weather conditions such as rainfall and high humidity.

[0061] Step S23: Calculate the ideal signal strength based on the base station's transmit power, obstruction loss, weather loss, base station antenna gain, and the antenna gain of each UAV in the low-altitude UAV group.

[0062] In some embodiments, such as Figure 3 As shown, step S23 includes steps S231 and S232.

[0063] Step S231, calculate the free space loss according to equation (2). .

[0064] (2)

[0065] In essence, free-space loss refers to the signal attenuation that occurs when radio waves propagate in a completely vacuum, unobstructed, ideal space due to the natural diffusion of wavefront energy. Physically, the electromagnetic waves radiated by a base station's transmitting antenna spread outwards like an expanding sphere. As the propagation distance increases, the same amount of energy is dispersed across an increasingly larger surface area, resulting in a decrease in energy density per unit area and consequently, a weaker signal captured by the receiving antenna. This loss is frequency-dependent—the higher the frequency (the shorter the wavelength), the greater the free-space loss over the same distance.

[0066] Step S232, calculate the ideal signal strength according to equation (3). .

[0067] (3)

[0068] In equation (3), It is the base station's transmission power. It is free space loss. It is weather loss. It's due to shading loss. It is the base station antenna gain. It is the antenna gain of each UAV in the low-altitude UAV group.

[0069] Understandably, Equation (3) quantifies all the key physical factors affecting signal propagation, including spatial diffusion, atmospheric absorption, obstacle blockage and antenna gain, forming a complete signal propagation loss model.

[0070] Understandably, through steps S21 to S23, the occlusion loss is calculated first, then the meteorological loss is calculated, and finally the ideal signal strength is calculated by combining the base station transmit power, occlusion loss, meteorological loss, base station antenna gain and UAV antenna gain. This provides a reliable benchmark value for subsequent signal compliance judgment, so that the judgment result no longer depends on a single factor, but integrates the three-dimensional environment, meteorological conditions and equipment parameters, thus improving the judgment accuracy.

[0071] Step S3: If the actual signal is determined to be substandard, and an interference source is identified, then: Control the mid-altitude UAV group to collect and transmit spectrum data of the area where the actual signal is substandard. Identify the type of interference source based on the spectrum data. Determine the interference handling plan based on the type of interference source. Control the low-altitude UAV group to locate the interference source and transmit its location coordinates. If the interference source is dynamic, control the low-altitude UAV group to track the interference source and transmit the tracking trajectory to obtain the movement trajectory of the interference source. Report the location or movement trajectory of the interference source and the interference handling plan.

[0072] For example, if low-altitude drones collect data... (Signal not up to standard) The medium-altitude UAV starts the spectrum data acquisition module to collect the signal spectrum data (sampling rate 1MSPS, spectrum resolution 1kHz) of the area (the area where the low-altitude UAV is located).

[0073] In some embodiments, in step S3, the mid-altitude UAV group is controlled to collect and transmit the spectrum data of the interference source. Identifying the type of interference source based on the spectrum data includes: controlling the mid-altitude UAV group to collect and transmit the spectrum data of the signal substandard area; extracting spectrum features from the spectrum data transmitted by the mid-altitude UAV group and analyzing the signal time series changes based on the Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) algorithm to determine the type of interference.

[0074] For example, a CNN+LSTM fusion algorithm is used to identify interference types: CNN layer: Extracts spectral features (such as bandwidth, peak power, and spectral shape). The input is a 256×256 spectral image, and the output is a 128-dimensional feature vector. LSTM layer: Analyzes the time series changes of signals (such as the period of periodic interference, the duration of sudden interference), the input is the feature vector output by CNN, and the output is the type of interference (such as WiFi interference, industrial electromagnetic interference, malicious interference).

[0075] Understandably, the fusion algorithm of convolutional neural networks and long short-term memory networks can accurately distinguish between wireless fidelity interference, industrial electromagnetic interference and malicious interference, with a significantly higher accuracy rate than the traditional threshold judgment method.

[0076] In some embodiments, step S3, controlling the low-altitude unmanned aerial vehicle (UAV) group to locate the interference source and transmit the location coordinates of the interference source, includes: scheduling three UAVs in the low-altitude UAV group to construct a triangulation network, and calculating and transmitting the three-dimensional coordinates of the interference source based on the angle of arrival and time difference of the interference signal transmitted back by each UAV in the triangulation network.

[0077] For example, three low-altitude drones are deployed to form a triangulation network, with each drone positioned at a distance from the interference source. Calculated using Angle of Arrival (AOA) and Time Difference of Arrival (TDOA): AOA measurement: The incident angle of the interference signal is measured using the directional antenna of the UAV (accuracy ±1°). ; TDOA measurement: Three drones simultaneously collect interference signals, calculate the signal arrival time difference (accuracy ±10ns), and combine this with the speed of light. ,get ; The coordinates of the interference source are solved using the least squares method. .

[0078] Understandably, triangulation networks can control the positioning error of interference sources within a small range. For example, in some examples, the positioning error is only eight meters, and they have the ability to dynamically track moving interference sources, thus solving the pain points of difficult positioning and tracking.

[0079] For example, if the interference source is a moving target (such as a ground jamming vehicle), the low-altitude UAV will activate dynamic tracking mode: Based on the changes in the intensity of the interference signal, adjust the flight speed (0-8m / s) to maintain a distance of ≤100m from the interference source; Record the trajectory of the interference source in real time (record coordinates once every 5 seconds) and generate a trajectory map (format: KML); For example, the method for generating and reporting a handling plan based on the type of interference can be as follows: If it is civilian interference such as WiFi / Bluetooth, a shutdown command is sent to the device to which the interference source belongs (such as a WiFi amplifier in a shopping mall) via an IoT module. If it is industrial electromagnetic interference, relevant departments are notified to conduct on-site investigation (such as poor grounding of factory equipment). If it is malicious interference, the low-altitude monitoring platform is coordinated to dispatch emergency drones to suppress the signal.

[0080] Through steps S1 to S3, the embodiments of this disclosure complete the layered deployment of the network, collaborative acquisition and calculation of ideal signal strength, determination of whether the signal meets the standard, identification of interference type if it does not meet the standard, location of interference source coordinates, tracking of moving interference sources and reporting of processing solutions. This achieves large-scale layered coverage, with three layers of UAVs operating collaboratively at different altitudes, significantly improving inspection efficiency; the inspection time for a 100-square-kilometer area is reduced from four hours to 1.2 hours. Simultaneously, it achieves a qualitative leap from simply knowing the signal is poor to knowing why it is poor, who is interfering, where the interference is occurring, and how to handle it. The entire process is completed through spectrum acquisition, artificial intelligence recognition, triangulation positioning, and dynamic tracking. Furthermore, the backend system only performs identification, location, solution generation, and reporting, without interfering with the underlying execution. The system architecture is clear and highly implementable.

[0081] In some embodiments, the method further includes: acquiring flight data and pedestrian flow data of the low-altitude unmanned aerial vehicle (UAV) group. The flight data, pedestrian flow data, and meteorological data of the low-altitude UAV group are input into a prediction model to obtain a predicted value of the signal strength within a preset future time period. The prediction model is based on historical data of the low-altitude UAV group, historical pedestrian flow data, historical meteorological data, and a long short-term memory network. If the predicted value of the signal strength in the first area within the preset future time period is lower than a preset threshold, then A is applied, using airway length, obstruction loss, and predicted signal strength as cost factors. The algorithm replans the UAV flight paths in the first region, obtaining an optimized flight path scheme, and then distributes the optimized flight path scheme to the relevant UAVs in the low-altitude UAV group, controlling the relevant UAVs to switch flight paths. The first region is a space area in the low-altitude layer.

[0082] For example, the preset time period can be several hours in the future (e.g., 6 to 20 hours in the future), and the prediction model is obtained as follows: The training data includes: Historical signal data: Low-altitude signal strength data for the past year (collected every 10 minutes, totaling 52,560 data points). Related data: meteorological data for the corresponding time period (obtained from China Weather Network API), pedestrian flow data (obtained from Gaode Map API), and low-altitude activity data (drone flight records); Model training: Input layer: 12 features (signal strength, temperature, humidity, rainfall, population density, number of drones, etc.); Hidden layers: 3 LSTM layers, 64 neurons per layer, using Dropout (dropout rate=0.2) to prevent overfitting; Output layer: Predicts signal strength for the next 6-24 hours (1 prediction value per hour); Model deployment: Deploy the trained model to the backend system, receive real-time data uploaded by multiple drones, and update the prediction results.

[0083] If the model predicts that the signal in a certain area will deteriorate within the next 6 hours (e.g., ... Will be lower Background startup path optimization: Based on a 3D environment model, using A The algorithm replans the drone's flight path: the cost function is ,in The length of the waterway is (km). ; For shading loss (dB). ; To predict signal strength (dBm). Prioritize routes with "low obstruction loss and high predicted signal strength," such as avoiding areas about to experience rainfall or densely populated areas with tall buildings; send the optimized routes to drones at all levels via dual-mode communication modules, and the drones will complete the route switching within 5 seconds to ensure uninterrupted inspections.

[0084] Understandably, the above method can achieve forward-looking predictions for six to twenty-four hours and actively avoid areas with deteriorating signals through dynamic path optimization, thereby significantly reducing the drone communication interruption rate and upgrading the system from passive response to proactive prevention.

[0085] In some embodiments, the method further includes: acquiring and recording the signal compliance rate and the coordinates of the weakest signal area in real time, and generating a signal quality report; acquiring and recording the type, quantity, location coordinates, and processing results of interference sources, and generating an interference source report; acquiring interference source processing reports and signal prediction reports for a preset future time period in real time, and generating a prediction report; generating and displaying a 3D signal heatmap, interference source trajectory animation, and signal prediction curve, and configuring corresponding data viewing permissions for different roles.

[0086] For example, the ground-based backend system receives signal acquisition data and interference source trajectory data uploaded by multiple UAVs and summarizes them in the following format: Signal quality report: including the signal compliance rate of each level and frequency band (e.g., the compliance rate of the 2.4G frequency band in the low-altitude layer is 92%), and the coordinates of the weakest signal area; Interference source report: including the type, number, location coordinates, and processing result of the interference source (e.g., "WiFi interference, turned off, processing time 15 minutes"); Prediction report: including the signal deterioration area in the next 24 hours and the optimized flight path plan.

[0087] For example, a visual interface can be generated via a web interface for visualization, supporting: 3D signal heatmap: Visually displays the signal intensity distribution at different heights and in different areas (red indicates weak signals, green indicates strong signals); Animation of the trajectory of the moving interference source: Plays the trajectory changes of the moving interference source; Prediction curve: Shows the trend of signal strength changes over the next 24 hours; Access control: Supports different roles such as low-altitude operation companies and regulatory authorities to view corresponding data.

[0088] Understandably, by transforming massive amounts of raw data into structured reports and intuitive visualization interfaces, different roles such as operating companies and regulatory authorities can quickly grasp the overall situation of the low-altitude wireless environment, providing convenient tools for decision-making and subsequent management, and further enhancing the practicality and scalability of this solution in actual engineering projects.

[0089] This disclosure presents an AI-based network optimization method that constructs a three-tiered UAV collaborative architecture (low-altitude, mid-altitude, and high-altitude) and integrates it with intelligent computing in the backend system. This enables comprehensive perception, accurate diagnosis, proactive prevention, and intuitive presentation of the low-altitude wireless environment. The low-altitude UAV is responsible for high-precision signal acquisition and 3D environment scanning; the mid-altitude UAV is responsible for spectrum data acquisition and data forwarding; and the high-altitude UAV is responsible for meteorological monitoring and data aggregation. These three systems are networked in a tree structure to transmit data back to the backend in real time. The backend calculates the ideal signal strength by comprehensively considering 3D environment modeling, obstacle classification and occlusion loss, meteorological loss, and free space loss, and compares this with measured values ​​to determine signal quality. When the signal is substandard, the backend accurately identifies the interference type using a convolutional neural network and long short-term memory network fusion algorithm, accurately locates the interference source coordinates using a triangulation network, tracks moving interference sources, generates a processing plan, and reports it. Furthermore, based on training the backend with historical signal, pedestrian flow, and meteorological data using the long short-term memory network, it can predict signal changes six to twenty-four hours in advance and, when signal deterioration is predicted, dynamically optimize the UAV's flight path using the A* algorithm to proactively avoid interference. The perception results are clearly presented through multi-dimensional reports and a 3D visualization interface. Compared with the existing passive detection mode of a single UAV with a fixed-gain antenna, this solution improves inspection efficiency by more than three times, significantly reduces the deviation between the ideal signal and the actual scene from 10 to 15 milliwatt decibels, achieves accurate location and dynamic tracking of interference sources, and significantly reduces communication interruption rate through advanced prediction and path optimization. It is suitable for various low-altitude scenarios such as urban core areas, logistics routes, and emergency rescue, and has strong versatility.

[0090] It is understood that the various method embodiments mentioned above in this disclosure can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this disclosure will not elaborate further. Those skilled in the art will understand that in the above methods of specific implementation, the specific execution order of each step should be determined by its function and possible internal logic.

[0091] In addition, this disclosure also provides a backend system, electronic device, and computer-readable storage medium, all of which can be used to implement any of the artificial intelligence-based network optimization methods provided in this disclosure. The corresponding technical solutions and descriptions are described in the relevant section of the method and will not be repeated here.

[0092] like Figure 4 As shown, this embodiment of the disclosure provides a background system 400, which includes a planning module 401, a signal receiving module 402, and an interference source processing module 403.

[0093] Planning module 401 is configured to: divide the inspection airspace into low-altitude, mid-altitude, and high-altitude layers based on the 3D map of the target area and according to altitude; and to deploy multiple inspection UAVs in a layered manner, configuring their network and flight path planning.

[0094] The signal receiving module 402 is configured to: control each of the multiple inspection drones to fly along its respective flight path, receive meteorological data from the high-altitude drone group, and lidar point cloud data and actual signal strength from the low-altitude drone group. It calculates the ideal signal strength based on the lidar point cloud data and meteorological data. It then compares the ideal signal strength with the actual signal strength to determine whether the actual signal meets the standard.

[0095] The interference source processing module 403 is configured as follows: If the actual signal is determined to be substandard, and an interference source is identified, then: the mid-altitude UAV group is controlled to collect and transmit the spectrum data of the area where the actual signal is substandard. The type of interference source is identified based on the spectrum data. An interference processing scheme is determined based on the type of interference source. The low-altitude UAV group is controlled to locate the interference source and transmit its location coordinates. If the interference source is dynamic, the low-altitude UAV group is controlled to track the interference source and transmit its tracking trajectory to obtain the movement trajectory of the interference source. The location or movement trajectory of the interference source and the interference processing scheme are then reported.

[0096] like Figure 5 As shown, this disclosure provides an electronic device, which includes: at least one processor 501; at least one memory 502; and one or more I / O interfaces 503 connected between the processor 501 and the memory 502; wherein the memory 502 stores one or more computer programs that can be executed by the at least one processor 501, and the one or more computer programs are executed by the at least one processor 501 to enable the at least one processor 501 to perform the above-described artificial intelligence-based network optimization method.

[0097] This disclosure also provides a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the aforementioned artificial intelligence-based network optimization method. The computer-readable storage medium may be volatile or non-volatile.

[0098] This disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device executes the above-described artificial intelligence-based network optimization method.

[0099] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0100] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0101] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0102] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0103] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0104] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0105] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0106] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0107] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0108] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in connection with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in connection with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this disclosure as set forth by the appended claims.

Claims

1. A network optimization method based on artificial intelligence, characterized in that, Applied to a backend system, the method includes: S1, based on the 3D map of the target area, divides the inspection airspace into low-altitude, mid-altitude, and high-altitude layers according to altitude; deploys multiple inspection drones in layers, configures their network, and plans their flight paths; S2, control each of the multiple inspection drones to fly according to its own flight path, receive meteorological data from the high-altitude drone group, lidar point cloud data and actual signal strength from the low-altitude drone group; calculate the ideal signal strength based on the lidar point cloud data and meteorological data; compare the ideal signal strength with the actual signal strength to determine whether the actual signal meets the standard. S3, if the actual signal is determined to be substandard and an interference source is identified, then: control the mid-altitude UAV group to collect and transmit the spectrum data of the area where the actual signal is substandard; identify the type of interference source based on the spectrum data; determine the interference handling scheme based on the type of interference source; control the low-altitude UAV group to locate the interference source and transmit the location coordinates of the interference source; if the interference source is a dynamic interference source, control the low-altitude UAV group to track the interference source and transmit the tracking trajectory to obtain the movement trajectory of the interference source; and report the location or movement trajectory of the interference source and the aforementioned interference handling scheme.

2. The network optimization method based on artificial intelligence according to claim 1, characterized in that, In S1, the layered deployment of multiple inspection drones includes: deploying low-altitude drone groups in the low-altitude layer, medium-altitude drone groups in the medium-altitude layer, and high-altitude drone groups in the high-altitude layer; wherein, each drone in the low-altitude drone group is equipped with a high-precision spectrum analyzer, lidar, and low-gain antenna unit; each drone in the medium-altitude drone group is equipped with a spectrum data acquisition module, a medium-gain antenna unit, and a dual-mode communication module; each drone in the high-altitude drone group is equipped with a long-range signal receiver, a multi-parameter meteorological sensor, and a high-gain antenna unit; the networking configuration of multiple inspection drones includes: using the high-altitude drone group as a core node, so that the high-altitude drone group receives data from the medium-altitude drone group and then transmits it back to the backend system; using the medium-altitude drone group as a relay node, so that the data collected by the low-altitude drone group is forwarded to the high-altitude drone group via the medium-altitude drone group.

3. The network optimization method based on artificial intelligence according to claim 2, characterized in that, In S2, methods for obtaining the ideal signal strength include: S21, calculate the occlusion loss based on the lidar point cloud data; S22, Calculate meteorological losses based on the meteorological data; S23. Calculate the ideal signal strength based on the base station's transmit power, obstruction loss, weather loss, base station antenna gain, and the antenna gain of each UAV in the low-altitude UAV group.

4. The network optimization method based on artificial intelligence according to claim 3, characterized in that, S21 includes: denoising the lidar point cloud data to obtain denoised point cloud data; using the ICP algorithm to obtain a three-dimensional mesh model based on the denoised point cloud data; classifying obstacles in the three-dimensional mesh model according to an artificial intelligence recognition model; and matching a preset dielectric constant to each obstacle according to its type. The thickness of each obstacle is calculated based on the three-dimensional mesh model. According to the formula Calculate the shading loss ,in, It is the wavelength of the signal emitted by the base station.

5. The network optimization method based on artificial intelligence according to claim 3, characterized in that, S22 includes: obtaining a meteorological loss coefficient based on the meteorological data and meteorological model, multiplying the meteorological loss coefficient by the actual transmission distance of the signal to obtain the meteorological loss; the actual transmission distance of the signal is the distance between the base station and the target drone, and the target drone is the drone to which the actual signal strength belongs.

6. The network optimization method based on artificial intelligence according to claim 3, characterized in that, S23 includes: S231, according to formula Calculate free space loss ; S232, according to formula Calculate the ideal signal strength ,in, It is the base station's transmission power. It is free space loss. It is weather loss. It's due to shading loss. It is the base station antenna gain. It is the antenna gain of each UAV in the low-altitude UAV group.

7. The network optimization method based on artificial intelligence according to claim 1, characterized in that, In S3 The controlled aerial drone group collects the spectrum data of the interference source and transmits it back; Identifying interference source types based on spectrum data includes: controlling a medium-altitude unmanned aerial vehicle (UAV) group to collect spectrum data in areas where the signal does not meet the standards and transmitting it back; and extracting spectrum features from the spectrum data transmitted back by the medium-altitude UAV group and analyzing the signal time series changes based on the Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) algorithm to determine the type of interference. The method of controlling the low-altitude unmanned aerial vehicle (UAV) group to locate the interference source and transmit the location coordinates of the interference source includes: scheduling three UAVs in the low-altitude UAV group to construct a triangulation network, and calculating and transmitting the three-dimensional coordinates of the interference source based on the angle of arrival and time difference of the interference signal transmitted back by each UAV in the triangulation network.

8. The network optimization method based on artificial intelligence according to claim 1, characterized in that, Also includes: Acquire flight data and pedestrian flow data of low-altitude unmanned aerial vehicle (UAV) groups; input the flight data, pedestrian flow data, and meteorological data of the low-altitude UAV groups into a prediction model to obtain the predicted signal strength within a preset future time period; the prediction model is based on historical data of low-altitude UAV groups, historical pedestrian flow data, historical meteorological data, and a long short-term memory network; if the predicted signal strength of the first region within the preset future time period is lower than a preset threshold, then A is used as a cost factor, with airway length, obstruction loss, and predicted signal strength as costs. The algorithm replans the UAV flight path in the first region to obtain an optimized flight path scheme, and then sends the optimized flight path scheme to the relevant UAVs in the low-altitude UAV group to control the relevant UAVs to switch flight paths; the first region is the spatial region in the low-altitude layer.

9. The network optimization method based on artificial intelligence according to claim 8, characterized in that, It also includes: real-time acquisition and recording of signal compliance rate and coordinates of the weakest signal area, generating a signal quality report; acquisition and recording of interference source type, quantity, location coordinates, and processing results, generating an interference source report; real-time acquisition of interference source processing reports and signal prediction reports for a preset future time period, generating a prediction report; generating and displaying a 3D signal heat map, interference source trajectory animation, and signal prediction curve, and configuring corresponding data viewing permissions for different roles.

10. A backend system, characterized in that, include: The planning module is set to divide the inspection airspace into low-altitude, mid-altitude, and high-altitude layers according to the height based on the 3D map of the target area. Layered deployment of multiple inspection drones, network configuration and flight path planning; The signal receiving module is configured to: control each of the multiple inspection drones to fly according to its own flight path, receive meteorological data from the high-altitude drone group, lidar point cloud data and actual signal strength from the low-altitude drone group; calculate the ideal signal strength based on the lidar point cloud data and meteorological data; and compare the ideal signal strength with the actual signal strength to determine whether the actual signal meets the standard. The interference source processing module is configured as follows: if the actual signal is determined to be substandard and an interference source is identified, then: control the mid-altitude UAV group to collect and transmit the spectrum data of the area where the actual signal is substandard; identify the type of interference source based on the spectrum data; determine the interference processing scheme based on the type of interference source; control the low-altitude UAV group to locate the interference source and transmit the location coordinates of the interference source; if the interference source is a dynamic interference source, control the low-altitude UAV group to track the interference source and transmit the tracking trajectory to obtain the movement trajectory of the interference source; and report the location or movement trajectory of the interference source and the aforementioned interference processing scheme.

11. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs that can be executed by the at least one processor, the one or more computer programs being executed by the at least one processor to enable the at least one processor to perform the artificial intelligence-based network optimization method as described in any one of claims 1-9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the artificial intelligence-based network optimization method as described in any one of claims 1-9.

13. A computer program product, characterized in that, Includes computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, wherein when the computer-readable code is run in a processor of an electronic device, the processor in the electronic device performs the artificial intelligence-based network optimization method as described in any one of claims 1-9.

Images