A mobile radio monitoring device layout scheduling and interference positioning method

Abstract

The present application relates to a kind of movable radio monitoring equipment layout scheduling and interference positioning method, belong to radio communication and spectrum monitoring technical field.This method constructs three-dimensional electromagnetic environment model, realizes monitoring equipment dynamic scheduling by improving genetic-reinforcement learning double algorithm and seven-dimensional adaptive fitness function;Adopt wavelet transform denoising to combine CNN-LSTM deep learning model to complete interference signal identification and screening;Use six-source fusion weighted positioning algorithm, cooperate scene adaptive propagation model, error correction and trajectory smoothing to improve positioning accuracy;While supporting data security transmission, multi-device collaborative operation and fault redundancy guarantee mechanism, form complete radio monitoring and interference positioning scheme.This method can adapt to complex geography and electromagnetic environment, effectively reduce monitoring blind area, improve interference identification and positioning accuracy, ensure that monitoring operation is continuous and stable, and can be used in spectrum management, public safety protection, emergency communication guarantee and other fields.

Description

Technical Field

[0001] This invention belongs to the fields of radio communication, intelligent spectrum monitoring, dynamic scheduling of multiple devices, and wireless signal positioning technology. Specifically, it relates to a method for deploying and scheduling mobile radio monitoring equipment and locating interference. This method is applicable to complex geographical and electromagnetic environments such as densely populated urban areas, mountainous terrain, forested hills, security for major events, emergency communication support, radio security in examination rooms, and the investigation and handling of illegal radio stations. It can achieve full-area radio signal monitoring, intelligent identification of illegal and interfering signals, and rapid and accurate location of interference sources. It is widely used in national radio spectrum management, public safety protection, emergency communication support, IoT spectrum supervision, aviation navigation radio safety, and railway dispatch communication protection, and is particularly suitable for spectrum monitoring scenarios that require dynamic adjustment of monitoring range, improved positioning accuracy, ensured monitoring continuity, and response to sudden interference. Background Technology

[0002] With the widespread adoption of 5G and 6G technologies, the Internet of Things (IoT), vehicle-to-everything (V2X) networks, industrial wireless control, and smart terminals, radio spectrum resources have become a critical national strategic resource, exacerbating the global spectrum supply-demand imbalance. Simultaneously, illegal activities such as unlicensed radio broadcasts, fake base stations, cheating devices used in exams, malicious interference transmitters, and unauthorized station setups are rampant, encroaching on legitimate spectrum resources, disrupting communication order, and interfering with critical services like aviation navigation, railway dispatching, emergency command, and public communications, threatening national security, public safety, and social stability.

[0003] Radio monitoring is a core technology for spectrum management. It uses specialized equipment to collect parameters such as signal frequency, amplitude, phase, modulation scheme, signal strength, time of arrival, and angle of arrival to perform signal identification, interference detection, and source localization. This is crucial for ensuring the legal use of the spectrum and investigating violations. Traditional monitoring systems, centered on fixed monitoring stations, offer 24 / 7 uninterrupted monitoring and high stability, but suffer from significant drawbacks: fixed coverage areas are affected by terrain and building obstructions, resulting in numerous monitoring blind spots in urban and mountainous regions; high deployment and maintenance costs make it difficult to achieve full coverage; and poor flexibility makes it unsuitable for dynamic monitoring needs such as temporary interference and sudden emergencies, leading to delayed interference location response.

[0004] Mobile radio monitoring equipment (vehicle-mounted, drones, portable terminals) is flexible, rapidly deployable, adaptable to complex terrain, and dynamically adjustable in location, making it a core component of modern spectrum monitoring systems. However, current technologies for the deployment, scheduling, and interference localization of mobile equipment still face numerous challenges.

[0005] 1. Insufficient intelligence and dynamism in layout and scheduling

[0006] Current scheduling relies heavily on manual experience and does not comprehensively consider constraints such as monitoring coverage, positioning accuracy, equipment energy consumption, mobility costs, communication links, traffic conditions, weather, and terrain obstruction. Some technologies use a single genetic algorithm or particle swarm optimization algorithm to complete the static initial layout, which cannot be dynamically adjusted in conjunction with real-time monitoring data, interference distribution, and equipment status. This easily leads to problems such as monitoring blind spots, resource waste, insufficient equipment endurance, and excessively high scheduling costs. Furthermore, a scheduling mechanism that combines proactive prediction and passive response has not been established, resulting in low efficiency in responding to sudden interference.

[0007] 2. Low accuracy in identifying interference signals and weak anti-interference capability.

[0008] Traditional identification relies solely on basic features such as signal strength and frequency, without deeply extracting multi-dimensional features such as time domain, frequency domain, modulation, and timing. This results in high false positive and false negative rates in scenarios with weak interference signals, multiple signal superposition, and complex electromagnetic noise. It is also unable to accurately distinguish between illegal broadcasts, malicious interference, unintentional interference, and stray interference from equipment, making it difficult to support the formulation of scheduling priorities and handling strategies, and resulting in low efficiency in interference investigation.

[0009] 3. The interference localization algorithm is singular, with poor robustness and environmental adaptability.

[0010] Current positioning methods mostly employ single AOA, TDOA, PDOA, or RSSI algorithms, each with its own shortcomings: AOA is greatly affected by multipath effects and occlusion, resulting in low positioning accuracy; TDOA has stringent time synchronization requirements, leading to large synchronization errors in dynamic movement scenarios; PDOA relies on a fixed propagation model, resulting in significant errors in complex terrain; RSSI is susceptible to environmental attenuation and noise interference, and is only suitable for open environments. Some multi-algorithm fusion technologies lack weighted optimization and do not incorporate NLOS error suppression, terrain occlusion correction, or trajectory smoothing mechanisms, resulting in significant positioning errors in complex urban environments and failing to meet the requirements for accurate investigation and detection.

[0011] 4. Environmental perception is disconnected from the physical environment, lacking the ability to perform dynamic modeling across the entire domain.

[0012] Existing technologies do not construct a digital model that integrates the physical and electromagnetic environments. They rely solely on two-dimensional plane maps for scheduling and positioning, failing to consider the impact of dynamic factors such as the three-dimensional structure of buildings, terrain elevation, dielectric loss, traffic accessibility, and real-time weather on signal propagation and equipment movement. As a result, scheduling and positioning are detached from the actual physical environment, which can easily lead to problems such as equipment being unable to reach the target location, incorrect signal model matching, and excessive positioning deviations, resulting in poor environmental adaptability.

[0013] 5. Lack of data processing, multi-device collaboration, and security mechanisms.

[0014] The massive amounts of monitoring data collected by mobile devices lack efficient noise reduction, compression, and transmission mechanisms, making them prone to data loss and transmission delays during network congestion, and compromising the real-time nature of core positioning data. Furthermore, the lack of clear division of labor and information sharing mechanisms among various types of mobile devices leads to low resource utilization due to independent operation. The absence of mechanisms for equipment fault redundancy, automatic return to base when battery is low, and backup equipment replacement results in insufficient monitoring continuity. Data transmission lacks high-strength encryption, posing risks of leakage and tampering, and failing to meet industry security compliance requirements.

[0015] Existing technologies only optimize the layout with a single algorithm, improve positioning accuracy by fusing multiple algorithms, or introduce digital twins to build environmental models. They have not achieved integrated applications such as multi-algorithm collaborative scheduling, deep learning interference identification, multi-source fusion positioning, 3D environment modeling, secure data transmission, and multi-device cluster collaboration. They have not formed a complete end-to-end solution and cannot fundamentally solve the core problems of mobile radio monitoring in complex scenarios. They are also unable to meet the actual needs of current spectrum management and interference investigation. Summary of the Invention

[0016] To address the problems of low intelligence in scheduling, poor interference identification accuracy, insufficient positioning robustness, weak environmental adaptability, inefficient data processing, lack of multi-device collaboration, and imperfect security of existing mobile radio monitoring equipment, this invention provides a method for the layout scheduling and interference localization of mobile radio monitoring equipment.

[0017] The technical objectives of this invention are as follows: to construct a three-dimensional electromagnetic environment model to achieve dynamic perception of the physical and electromagnetic environments across the entire domain; to adopt an improved genetic-reinforcement learning dual-algorithm collaborative scheduling mechanism, combined with a seven-dimensional adaptive fitness function, to achieve active prediction and dynamic response scheduling of mobile devices; to introduce a CNN-LSTM deep learning model, coupled with wavelet transform denoising preprocessing, to achieve high-precision identification and classification of interference signals; to propose a six-source fusion weighted positioning algorithm, integrating the advantages of multiple algorithms, scene adaptive propagation model, NLOS suppression, terrain correction, and trajectory smoothing, to achieve accurate positioning of interference sources in complex environments; and to improve data compression, encryption, congestion control, multi-device cluster collaboration, and fault redundancy mechanisms to ensure data security, monitoring continuity, and efficient equipment collaboration, thereby achieving full-scene radio monitoring and interference positioning.

[0018] The method of this invention relies on four hardware modules: a central control platform, a cluster of mobile monitoring devices, a hybrid data transmission module, and a positioning and interaction terminal. It is supported by seven software modules: environmental modeling, scheduling decision-making, interference identification, positioning calculation, data security, closed-loop control, and equipment management, which work together to form a complete intelligent monitoring and positioning system.

[0019] Hardware module composition

[0020] 1. Central Control Platform

[0021] It adopts an industrial-grade high-performance server cluster, configured with Intel Xeon series processors, 128GB or more of memory, 4TB or more of high-speed solid-state drives, and runs on Windows Server 2019 / Linux CentOS 7.0 operating system. It has the capabilities of massive data processing, high-speed algorithm operation, and real-time command issuance. Its core functions include 3D environment modeling, task parsing, equipment status perception, scheduling algorithm operation, interference identification, location calculation, data storage, command issuance, status monitoring, and report generation.

[0022] 2. Mobile monitoring equipment cluster

[0023] By employing a heterogeneous combination of vehicle-mounted, drone-based, and portable monitoring terminals, an integrated air-ground monitoring system is constructed, with each type of equipment performing complementary functions.

[0024] ① Vehicle-mounted monitoring terminal: Equipped with a high-gain omnidirectional antenna, a high-power SDR software radio receiver, a GPS / BeiDou dual-mode timing and positioning module, and a large-capacity lithium battery, with a battery life of ≥8 hours and a coverage radius of ≥5km. It has high receiving sensitivity and is responsible for monitoring and positioning anchor point support in large open areas and along roads.

[0025] ②UAV monitoring terminal: It adopts a multi-rotor platform and is equipped with a lightweight SDR receiver, RTK high-precision positioning module and miniature direction-finding antenna. It is mobile and flexible and can monitor at high altitudes without obstruction. It is responsible for blind spot monitoring and accurate direction finding in densely built-up urban areas, mountainous ravines and other complex terrains.

[0026] ③ Portable monitoring terminal: Weighing ≤2kg, easy to hold, equipped with a directional antenna, low-power SDR receiver, and GPS positioning module, responsible for close-range investigation of interference sources, on-site evidence collection and accurate positioning, and suitable for emergency response and indoor interference search.

[0027] 3. Hybrid data transmission module

[0028] Employing a three-tiered hybrid communication architecture—primarily wired, supplemented by 5G, and further enhanced by a mesh self-organizing network—this architecture ensures seamless data transmission between devices and the central control platform across all scenarios. The wired network guarantees stable transmission for vehicle-mounted terminals, the 5G network ensures real-time mobile transmission for drones and portable terminals, and the mesh self-organizing network guarantees connectivity in public network blind spots and emergency scenarios. The module features data compression, encryption, and congestion control functions to ensure efficient, secure, and stable transmission.

[0029] 4. Positioning Interactive Terminal

[0030] Employing industrial-grade tablet PCs and industrial control computers, and equipped with visual interactive software, it can display the location of interference sources, monitor equipment status, render 3D environmental maps, query monitoring data, and support manual command intervention, providing operational guidance for on-site interference troubleshooting.

[0031] Software module composition

[0032] 1. Environment Modeling Module: Completes 3D grid division, grid attribute assignment, dynamic environment map construction and real-time updating;

[0033] 2. Scheduling Decision Module: Completes the operation of the improved genetic-reinforcement learning dual algorithm, the calculation of the seven-dimensional fitness function, dynamic delay scheduling, and the generation of scheduling schemes;

[0034] 3. Interference Identification Module: Performs wavelet transform denoising, multidimensional signal feature extraction, CNN-LSTM model identification, interference screening and classification;

[0035] 4. Positioning and calculation module: Completes time synchronization calibration, preliminary positioning using multiple algorithms, scene adaptive propagation model matching, six-source fusion weighting, EKF trajectory smoothing, and accuracy verification;

[0036] 5. Data Security Module: Completes data compression, AES-256 encryption, multi-factor congestion control, data storage and traceability;

[0037] 6. Closed-loop control module: Completes positioning result evaluation, GDOP optimization, rescheduling decision, and disturbance escape prediction;

[0038] 7. Equipment Management Module: Completes equipment status monitoring, low power warning, fault detection, backup equipment scheduling, and energy consumption statistics.

[0039] The method of this invention consists of eight core steps, which are logically coherent and executed in a closed loop, realizing intelligent operation of the entire process from environmental perception to interference handling. The specific solution is as follows:

[0040] Step S1: 3D Digital Modeling of Dynamic Electromagnetic Environment

[0041] The central control platform acquires information such as topography, 3D building structure, road distribution, traffic flow, real-time weather, historical spectrum data, and interference event records in the monitored area through GIS, urban BIM, DEM digital elevation model, traffic big data, and meteorological platform. It employs an adaptive octree grid partitioning algorithm to complete 3D spatial meshing: the grid resolution is set to 5m×5m×5m in densely built-up areas and areas with high historical interference incidence to improve environmental perception accuracy; the resolution is set to 50m×50m×50m in open areas and non-sensitive mountainous areas to reduce algorithm computation. Each 3D grid is assigned six core attributes: terrain occlusion coefficient, dielectric loss coefficient, background noise level, historical interference probability, traffic accessibility, and scene type label (urban / mountainous / open). The 3D grid is then integrated with real-time spectrum data, equipment status, and interference distribution to generate a dynamic electromagnetic environment map, achieving real-time mapping and visualization of the physical world, electromagnetic environment, and equipment status, providing environmental data support for scheduling and positioning.

[0042] Step S2: Monitoring Task Analysis and Device Resource Status Awareness

[0043] Task Analysis: The central control platform receives tasks such as sudden interference alarms, routine inspections, major event support, and emergency communication monitoring. It automatically analyzes core parameters such as target monitoring frequency band, interference type, task priority, positioning accuracy requirements, monitoring duration, and coverage area. Resource Awareness: Through the hybrid data transmission module, it collects information such as the real-time location, remaining battery power, battery life threshold, device sensitivity, communication link status, task load, and fault status of the mobile device cluster. Devices with less than 20% remaining battery power, communication interruption, or hardware failure are marked as unavailable and excluded from the scheduling candidate set to ensure the execution of the scheduling plan.

[0044] Step S3: Improve the dynamic layout scheduling of the genetic-reinforcement learning dual-algorithm collaboration.

[0045] This step integrates an improved genetic algorithm (adaptive mutation probability), a multi-factor dynamic fitness function, dynamic delayed scheduling and reinforcement learning scheduling, and a disturbance risk potential field to form a dual-algorithm collaborative scheduling mechanism. This avoids the problems of a single algorithm getting stuck in local optima and having poor dynamic adaptability. Specific implementation details are as follows:

[0046] 1. Construct a seven-dimensional adaptive fitness function

[0047] A dynamic fitness function is established based on seven factors: comprehensive monitoring coverage, positioning accuracy, equipment energy consumption, mobility cost, communication loss, interference risk, and traffic accessibility: F = αS + βP - γE - δC - εL - ζR - ηT. Here, F is the fitness value, with a higher value indicating a better scheduling scheme; S is the monitoring coverage area percentage; P is the positioning accuracy compliance rate; E is the total equipment energy consumption; C is the total equipment mobility cost; L is the communication link loss; R is the interference risk potential field value; and T is the traffic accessibility coefficient. α, β, γ, δ, ε, ζ, and η are weighting coefficients, summing to 1. The weights are dynamically adjusted according to task priority (increased weights for sudden interference and increased weights for β and ζ, and increased weights for routine inspections).

[0048] 2. Improved collaborative optimization of genetic-reinforcement learning dual algorithms

[0049] ① Initial layout optimization: An improved genetic algorithm is adopted, using device location coordinates as chromosomes and real-number encoding, introducing adaptive mutation probability, and dynamically adjusting the mutation probability according to the population fitness variance to avoid local optima and improve convergence speed. The optimal initial layout is generated through selection, crossover, and mutation iteration. ② Dynamic scheduling decision: A reinforcement learning algorithm is adopted, using a dynamic electromagnetic environment map as input and a seven-dimensional fitness function as the reward function. Combined with the interference risk potential field, the algorithm guides the device to prioritize covering high-risk areas, realizing dynamic scheduling autonomous decision-making. ③ Dual algorithm collaboration: The improved genetic algorithm is responsible for static initial layout, and reinforcement learning is responsible for dynamic real-time adjustment, realizing full-cycle scheduling that combines static optimization and dynamic adaptation.

[0050] 3. Dynamic delay scheduling execution

[0051] The scheduling waiting threshold is dynamically calculated based on task type, equipment status, network congestion level, and traffic conditions. Triggering conditions include: monitoring coverage blind spot ratio > 5%, remaining equipment battery power < battery life threshold, interference signals detected in key areas, sudden traffic / weather changes affecting equipment movement, and substandard positioning accuracy. Based on the triggering conditions, the central control platform generates an optimal scheduling plan including movement path, target location, monitoring frequency band, and dwell time, and distributes it to each device.

[0052] Step S4: Multi-device collaborative data acquisition and wavelet transform preprocessing

[0053] Each mobile device arrives at the target location according to the dispatch instructions, starts full-band monitoring, and collects signal parameters such as center frequency, signal amplitude, phase, modulation mode, RSSI, TOA, and AOA in real time. It also collects its own location and timing information through the GPS / BeiDou dual-mode module, and binds the signal parameters with location and time to form structured data.

[0054] Data preprocessing employs a wavelet transform denoising algorithm, which decomposes the signal into multiple layers using wavelet basis functions to separate environmental noise, equipment noise, and valid signals. After noise removal, the signal is reconstructed, improving the signal-to-noise ratio. The reconstructed signal is then normalized to the [0,1] interval, providing standardized data for interference identification. Preprocessed data is uploaded via a hybrid transmission module, prioritizing the transmission of core positioning data such as TOA, AOA, and RSSI.

[0055] Step S5: CNN-LSTM Deep Learning Interference Identification and Precise Filtering

[0056] This step uses a CNN-LSTM deep learning model to replace traditional basic feature recognition, achieving high-precision identification and classification of interference signals:

[0057] 1. Multidimensional feature extraction: Extract four dimensions of features from the preprocessed signal: time domain (peak value, mean, variance, pulse width, duration), frequency domain (spectral peak value, bandwidth, harmonic components, occupancy), modulation (modulation method, depth, bit error rate), and time series (signal strength variation, intermittent transmission), and integrate them into a high-dimensional feature vector as model input;

[0058] 2. CNN-LSTM Model Recognition: The model consists of convolutional layers, recurrent layers, and fully connected layers. Convolutional layers extract static time-domain / frequency-domain features of the signal, recurrent layers capture temporal variation features and adapt to intermittent interference, and fully connected layers complete signal classification, outputting legitimate / interference signals and specific interference types. The model is trained on massive amounts of labeled data, achieving a recognition accuracy of ≥98.5%, and a recognition rate of ≥95% for weak interference signals (≤-75dBm).

[0059] 3. Interference signal screening: Valid interference sources are screened based on preset thresholds, namely, signal strength > -80dBm, continuous transmission time > 5 minutes, and frequency not in the legal spectrum library. These are marked as interference sources to be located and relevant information is recorded before entering the positioning process.

[0060] Step S6: Six-source fusion weighted interference localization and accuracy optimization

[0061] This step integrates AOA / TDOA / PDOA weighted algorithms, scene adaptive propagation model, RSSI constraint, NLOS error suppression, terrain occlusion correction, and EKF trajectory smoothing to form a six-source fusion weighted localization algorithm, achieving accurate localization in complex environments:

[0062] 1. Positioning device selection and time synchronization: Select ≥3 devices with excellent signal reception quality, reasonable location distribution, and low GDOP as cooperative positioning devices, and give priority to devices close to suspected interference sources; adopt GPS time synchronization + NTP network protocol dual calibration to ensure time error ≤1μs to meet high-precision synchronization requirements;

[0063] 2. Scene Adaptive Propagation Model Matching: Based on the scene labels of the dynamic electromagnetic environment map, the COST231-Hata model is matched for urban building areas, the Okumura-Hata model is matched for mountainous and hilly areas, and the free space propagation model is matched for open areas.

[0064] 3. Preliminary localization using multiple algorithms: The preliminary coordinates of the interference source are calculated using AOA, TDOA, and PDOA algorithms respectively;

[0065] 4. Six-source fusion weighted solution: It integrates data from six dimensions: AOA, TDOA, PDOA, RSSI constraint, NLOS suppression, and terrain correction. The final coordinates are calculated through a weighted fusion algorithm, and the weight coefficients are dynamically allocated. The higher the positioning accuracy and the stronger the environmental adaptability, the greater the weight of the dimension.

[0066] 5. EKF trajectory smoothing and accuracy verification: Extended Kalman filter is used to smooth the positioning trajectory to eliminate measurement noise and abrupt errors; the deviation between the final coordinates and the initial coordinates is calculated. If the deviation is ≤10 meters, the result is valid. If it exceeds the standard, the equipment is reselected for positioning until the accuracy meets the standard.

[0067] Step S7: Data Transmission Optimization and End-to-End Security Assurance

[0068] This step addresses congestion and security risks associated with massive data transmission through data compression, encrypted transmission, and multi-factor congestion control.

[0069] 1. Data Compression: Employing the LZ77 lossless compression algorithm, data transmission volume is reduced by more than 60%, alleviating network load;

[0070] 2. Encrypted transmission: Data and instructions are encrypted using the AES-256 encryption algorithm to prevent tampering, theft, and leakage, in accordance with national spectrum monitoring data security standards;

[0071] 3. Multi-factor congestion control: Dynamically adjust the upload rate according to network congestion level, data type, and device status. Reduce the regular frequency sweep data rate during severe congestion to prioritize core data transmission.

[0072] 5. Data Management: The central control platform stores monitoring, positioning, and scheduling data in real time, supports tracing, querying, and statistics, and automatically generates spectrum monitoring reports.

[0073] Step S8: Closed-loop optimization and multi-device fault redundancy protection

[0074] 1. Location result feedback and handling guidance: Push the coordinates, type, and signal parameters of the interference source to the positioning interactive terminal, and combine the location marked on the dynamic electromagnetic environment map to guide law enforcement personnel to conduct on-site investigation and evidence collection.

[0075] 2. Closed-loop rescheduling optimization: When the positioning accuracy is not up to standard, rescheduling is automatically triggered to direct the nearest equipment to approach the center of the error ellipse and optimize the GDOP factor; when the interference source emits intermittently or escapes, the escape path is predicted and the equipment is scheduled to set up ambush for monitoring.

[0076] 3. Multi-device redundancy: Real-time monitoring of equipment status, automatic return to charging of low-battery equipment, and replacement of faulty equipment by backup equipment to ensure uninterrupted monitoring; clear division of labor among vehicle-mounted, drone, and portable equipment to form a collaborative system for large-area coverage, blind spot filling in complex terrain, and close-range inspection.

[0077] Beneficial effects

[0078] This invention combines multi-algorithm optimization, deep learning recognition, 3D environment perception, multi-source fusion localization, and end-to-end security assurance to specifically address the core problems of existing technologies. Specific beneficial effects are as follows:

[0079] 1. Improve the intelligence level of scheduling: By improving the genetic-reinforcement learning dual-algorithm collaborative scheduling and the seven-dimensional adaptive fitness function, the dual-mode scheduling of active prediction and passive response is realized. Taking into account multiple constraints, the monitoring coverage blind spot is controlled within 3.5%; the integrated air-space-ground equipment cluster collaborative operation improves resource utilization by more than 40%, significantly reduces equipment energy consumption and scheduling costs, and greatly shortens the scheduling response time for sudden interference.

[0080] 2. Improve interference identification accuracy: By adopting wavelet transform denoising and CNN-LSTM deep learning model, multi-dimensional features of the signal are extracted in depth, the weak interference signal identification rate is ≥95%, and the interference identification accuracy is ≥98.5%; it can accurately distinguish the interference type, provide data support for scheduling priority and handling strategy, reduce the false judgment and false miss rate, and improve the efficiency of interference investigation.

[0081] 3. Enhanced positioning robustness: The six-source fusion weighted positioning algorithm integrates the advantages of multiple algorithms, and is equipped with a scene adaptive propagation model, NLOS error suppression, terrain occlusion correction, and EKF trajectory smoothing mechanism. The positioning error in complex urban environments is ≤20 meters, and the line-of-sight environment error is ≤1.5 meters. It overcomes the shortcomings of single algorithms and improves positioning robustness and environmental adaptability.

[0082] 4. Achieve precise environmental perception: Construct a dynamic three-dimensional electromagnetic environment model, integrate physical elements such as terrain, buildings, transportation, and meteorology with electromagnetic data, realize real-time mapping between the physical world and the electromagnetic environment, and ensure that scheduling and positioning are in line with the actual scene, avoiding problems such as equipment being unable to reach, model matching errors, and excessive positioning deviations.

[0083] 5. Enhanced Data and Collaboration Support: Through LZ77 data compression, AES-256 encryption, and multi-factor congestion control, the data transmission success rate is ≥99.95%, with no risk of data loss or leakage; a multi-device division of labor and collaboration mechanism and fault redundancy mechanism are established, with automatic return to base when low battery and automatic fault replacement, ensuring 100% monitoring continuity and meeting the stable monitoring needs of all scenarios. Attached Figure Description

[0084] Figure 1 This is a schematic diagram of the overall system architecture of the present invention;

[0085] Figure 2 This is a flowchart illustrating the overall process of the method of the present invention.

[0086] Figure 3 This is a detailed flowchart of the improved genetic-reinforcement learning dual-algorithm scheduling method of the present invention;

[0087] Figure 4 This is a detailed flowchart of the six-source fusion weighted interference localization process of the present invention. Detailed Implementation

[0088] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. These embodiments are only used to explain the present invention and are not intended to limit the scope of protection. All embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present invention.

[0089] 1. Implement system hardware configuration

[0090] This embodiment selects a complex monitoring area of ​​50 km² in the city (including densely built-up areas, low hills, and open roads), and the hardware configuration is as follows:

[0091] (1) Central control platform: 2 industrial-grade servers, Intel Xeon E5-2699v4 processor, 128GB memory, 4TB SSD, and seven software modules deployed.

[0092] (2) Mobile monitoring equipment cluster: 3 vehicle-mounted terminals, 4 UAV terminals and 2 portable terminals, all equipped with GPS / BeiDou timing and positioning modules, supporting monitoring in the 100kHz-3GHz frequency band;

[0093] (3) Hybrid data transmission module: wired gigabit network, 5G mobile communication, Mesh self-organizing network, equipped with AES-256 encryption, LZ77 compression, and multi-factor congestion control functions;

[0094] (4) Positioning and interaction terminals: 5 industrial tablet computers equipped with visual interaction software.

[0095] 2. Algorithm parameter configuration

[0096] (1) Improved genetic algorithm: population size 20, number of iterations 100, crossover probability 0.8, adaptive mutation probability 0.05-0.2;

[0097] (2) Reinforcement learning: The reward function is a seven-dimensional fitness function, the learning rate is 0.001, and the number of iterations is 200;

[0098] (3) CNN-LSTM model: 2 convolutional layers, 2 LSTM recurrent layers, 1 fully connected layer, 50 training iterations, 0.001 learning rate, 10,000 training data sets;

[0099] (4) Positioning parameters: time synchronization error ≤1μs, positioning error threshold 10 meters, number of cooperative positioning devices ≥3;

[0100] (5) Scheduling parameters: coverage blind zone threshold 5%, equipment endurance threshold 20%, scheduling cycle 10 minutes.

[0101] 3. Specific Implementation Steps

[0102] Step S1: The central control platform for dynamic electromagnetic environment 3D modeling acquires GIS, BIM, DEM, traffic, and meteorological data of the monitoring area. It adopts an adaptive octree grid division, with a resolution of 5m×5m×5m in densely built areas and 50m×50m×50m in open areas. Each grid is assigned a terrain occlusion coefficient, dielectric loss coefficient, background noise, historical interference probability, traffic accessibility, and scene type label to construct a dynamic electromagnetic environment map and map the regional environment and electromagnetic state in real time.

[0103] Step S2: Task analysis and device perception system receives illegal broadcast interference to investigate sudden tasks. The target frequency band is 87-108MHz, with the highest priority and a positioning accuracy requirement of ≤10 meters. Real-time collection of device status: all devices have a remaining power of ≥30%, no faults, and normal communication, and are all included in the scheduling candidate set.

[0104] Step S3: Dual-algorithm cooperative scheduling

[0105] (1) Weight configuration of the seven-dimensional fitness function: α=0.2, β=0.25, γ=0.15, δ=0.1, ε=0.1, ζ=0.15, η=0.05;

[0106] (2) Improve the genetic algorithm to complete the initial layout: 3 vehicle-mounted terminals are deployed on open roads in the east, west and north of the area, 4 drones are deployed around the densely built-up area, and 2 portable terminals are used as backups;

[0107] (3) Enhance learning and dynamic optimization: Real-time monitoring of signal distribution, detection of signal anomalies in the southeast building area, dispatching 1 drone and 1 portable terminal to move to the area to optimize monitoring coverage;

[0108] (4) Dynamic delay scheduling: Without triggering conditions, the equipment executes according to the plan, and the monitoring coverage blind area accounts for 3.2%, which meets the requirements.

[0109] Step S4: Data Acquisition and Wavelet Preprocessing. Each device acquires signal parameters according to instructions. The wavelet transform algorithm removes environmental noise and equipment noise. After normalization, the data is encrypted and uploaded through the hybrid transmission module to prioritize the transmission of positioning data and ensure network congestion-free operation.

[0110] Step S5: CNN-LSTM Interference Identification

[0111] (1) Extract the time domain, frequency domain, modulation, and timing features of the signal to generate a high-dimensional feature vector;

[0112] (2) Input the trained CNN-LSTM model to identify the 95MHz signal in the southeast building area as illegal broadcast interference, with a signal strength of -55dBm and a duration of 15 minutes;

[0113] (3) Select the effective interference sources to be located and record the signal parameters and equipment information.

[0114] Step S6: Six-Source Fusion Positioning

[0115] (1) Select one vehicle-mounted device, one drone, and one portable terminal as collaborative positioning devices, with GPS+NTP time synchronization and an error of 0.5μs;

[0116] (2) The scenario is a densely populated urban building area, which is matched with the COST231-Hata propagation model;

[0117] (3) Calculate the preliminary coordinates of AOA, TDOA and PDOA respectively, and integrate RSSI constraints, NLOS suppression and terrain correction to calculate the final coordinates using a six-source weighted solution;

[0118] (4) After EKF smoothing, the positioning error is 2.1 meters, which is ≤10 meters threshold, and the positioning result is valid.

[0119] Step S7: Secure Data Transmission. Data is uploaded after LZ77 compression and AES-256 encryption, with a 100% success rate and no leakage or tampering. The system automatically stores the data and generates an interference monitoring report.

[0120] Step S8: Closed-loop processing and redundancy protection

[0121] (1) The location results were pushed to the interactive terminal, indicating that the interference source was located on the roof of a residential building. Law enforcement officers investigated and dealt with the matter on-site and dismantled the illegal transmitting equipment;

[0122] (2) After the interference is eliminated, the system scheduling equipment is restored to its initial layout, and the low-battery UAV automatically returns to its home base to recharge.

[0123] (3) The monitoring was uninterrupted throughout the process, and the equipment worked together smoothly to complete the interference investigation task.

[0124] 4. Performance Verification and Comparison Experiments

[0125] This embodiment compares the method of the present invention with traditional manual scheduling + single TDOA positioning and conventional mobile monitoring solutions in a 72-hour comparative experiment, simulating 10 interference events (5 illegal broadcasts, 3 malicious interferences, and 2 unintentional interferences). The core performance indicators are compared as follows:

[0126] sheet

[0127] Performance indicators Method of the present invention Conventional mobile monitoring solutions Traditional methods Percentage of monitoring blind spots 2.8%-3.5% 3.5%-4.0% 6.2%-8.7% Interference identification accuracy 98.5% 95.0% 81.5% Urban NLOS positioning error ≤20 meters ≤30 meters ≥50 meters Line-of-sight positioning error ≤1.5 meters ≤2.0 meters ≥15 meters Scheduling response time ≤2 minutes ≤2.5 minutes ≥30 minutes Equipment resource utilization rate ≥92% 85.0%-88.0% 51.2%-58.7% Data transmission success rate ≥99.95% 99.7% 95.0%

[0128] Experimental results show that the method of the present invention is superior to the existing technology in terms of identification accuracy, positioning robustness, scheduling efficiency, resource utilization and data security, and can meet the actual needs of radio monitoring and interference investigation in complex scenarios.

[0129] 5. Multi-scenario adaptation testing

[0130] The method of this invention was tested in four typical scenarios: densely built-up urban areas, mountainous and hilly areas, open areas, and major event venues. The performance results are as follows:

[0131] In densely populated urban areas: coverage blind spots are 3.2%, positioning error is 2.5 meters, and recognition accuracy is 97.8%.

[0132] Mountainous and hilly areas: Coverage blind spots of 3.5%, positioning error of 2.8 meters, and recognition accuracy of 97.5%;

[0133] Open area: Coverage of blind spots is 2.8%, positioning error is 1.5 meters, and recognition accuracy is 98.7%;

[0134] At major events: multi-target parallel monitoring with a response time of ≤1.5 minutes and no interference-induced missed detections.

[0135] 6. Fault and Extreme Scenario Handling

[0136] (1) Low battery: When the battery level is less than 20%, the system will automatically schedule the return trip for charging and the backup equipment will be used to replace it immediately;

[0137] (2) Communication interruption: When the public network is interrupted, the Mesh self-organizing network will be automatically switched to ensure data transmission;

[0138] (3) Severe NLOS obstruction: Schedule UAVs to take off and acquire line-of-sight signals to optimize positioning accuracy;

[0139] (4) Interference escape: The system predicts the escape path, schedules equipment to set up an ambush, and locates the source immediately after the signal is reproduced.

[0140] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention. The actual structure is not limited thereto. Any structural methods and embodiments similar to this technical solution devised by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope of the present invention.

Claims

1. A method for the layout scheduling and interference localization of mobile radio monitoring equipment, characterized in that: The process is carried out by a central control platform, a cluster of mobile monitoring devices, a hybrid data transmission module, and a positioning and interaction terminal, and includes the following steps: S1 Dynamic Electromagnetic Environment 3D Modeling: Integrating geographic information and electromagnetic data, completing 3D grid division and attribute assignment, and generating a dynamic electromagnetic environment map; S2 Task Analysis and Equipment Status Awareness: Analyze the core parameters of the monitoring task, collect equipment operating status, and filter available equipment; S3 Improved Genetic-Reinforcement Learning Dual-Algorithm Cooperative Scheduling: Based on a seven-dimensional adaptive fitness function, it completes the initial layout and dynamic scheduling of equipment through dual algorithms and generates a scheduling scheme; S4 Data Acquisition and Wavelet Preprocessing: The equipment acquires signals and spatiotemporal data according to the scheduling instructions, and uploads them after wavelet denoising and normalization processing; S5 Interference Identification and Screening: Extract multi-dimensional features of the signal, classify them using a model, and then screen for effective interference sources; S6 Interference Localization and Accuracy Optimization: Select a cooperative localization device and complete time synchronization, match the scene adaptation propagation model, and obtain the localization result through multi-source fusion weighting and trajectory smoothing; S7 Data Transmission and Security: Stable and secure data transmission is achieved through data compression, encryption, and congestion control. S8 Closed-Loop Optimization and Equipment Redundancy Assurance: Based on the positioning results, closed-loop rescheduling is performed, and a low-power and fault redundancy mechanism for equipment is established to ensure continuous monitoring.

2. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: Step S1 employs an adaptive octree grid partitioning algorithm, with a grid resolution of 5m×5m×5m for densely built-up areas and areas with high historical interference rates, and 50m×50m×50m for open areas and non-sensitive mountainous areas. The three-dimensional grid is then labeled with terrain occlusion coefficient, dielectric loss coefficient, background noise level, historical interference probability, traffic accessibility, and scene type. Combined with GIS, BIM, DEM, traffic, meteorological, and historical spectrum data, a dynamic electromagnetic environment map is generated.

3. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: The seven-dimensional adaptive fitness function described in step S3 is constructed by comprehensively monitoring coverage, positioning accuracy, equipment energy consumption, mobility costs, communication loss, interference risk, and traffic accessibility, with weights dynamically adjusted according to task priority. The improved genetic algorithm uses equipment location coordinates as chromosomes and employs real-number encoding and adaptive mutation probability to complete the initial layout. Reinforcement learning uses a dynamic electromagnetic environment map as input and the seven-dimensional adaptive fitness function as the reward function to complete dynamic scheduling. Dynamic delay scheduling is triggered by exceeding coverage blind spots, insufficient equipment power, detection of interference signals, sudden environmental changes, and substandard positioning accuracy.

4. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: In step S4, the mobile monitoring device collects center frequency, signal amplitude, phase, modulation mode, RSSI, TOA, and AOA parameters, and combines them with GPS / BeiDou timing and positioning information to form structured data; a wavelet transform denoising algorithm is used to separate noise from effective signals and complete normalization; the hybrid data transmission module adopts a three-level communication architecture of wired, 5G, and Mesh self-organizing network, prioritizing the transmission of core positioning data.

5. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: Step S5 extracts the time domain, frequency domain, modulation, and timing dimension features of the signal, inputs them into the CNN-LSTM model to complete the identification and classification; and filters valid interference sources according to signal strength, duration, and spectrum legality threshold.

6. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: Step S6: Select no fewer than 3 devices with good signal quality, reasonable location distribution, and low GDOP value as collaborative positioning devices, and use GPS timing and NTP network protocol for dual time calibration. Based on the scene label matching propagation model, the COST231-Hata model is used in urban building areas, the Okumura-Hata model is used in mountainous and hilly areas, and the free space propagation model is used in open areas. The trajectory is smoothed by integrating AOA, TDOA, PDOA, RSSI constraints, NLOS error suppression, and terrain occlusion correction data and then applying extended Kalman filtering.

7. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: Step S7 uses the LZ77 lossless compression algorithm to reduce the amount of data, and uses the AES-256 encryption algorithm to ensure data security. Congestion control is implemented based on network congestion level, data type, and device status. The central control platform completes data storage, traceability, and monitoring report generation.

8. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: The mobile monitoring equipment cluster is composed of vehicle-mounted, drone-based, and portable terminals in a heterogeneous configuration. Vehicle-mounted terminals are used for monitoring and positioning anchor point support in large open areas, UAV terminals are used for blind spot detection and direction finding in complex terrain, and portable terminals are used for close-range investigation and evidence collection, forming an integrated air-space-ground monitoring system.

9. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: In step S8, when the positioning accuracy is not up to standard, rescheduling is automatically triggered to optimize the GDOP factor; when the interference source emits intermittently or moves, the path is predicted and the equipment is deployed for monitoring; the positioning result is pushed to the positioning interactive terminal, and the location is marked in combination with the three-dimensional electromagnetic map to guide on-site handling.

10. The method for layout scheduling and interference localization of mobile radio monitoring equipment according to claim 1, characterized in that: Step S8 monitors the equipment status in real time. Equipment with remaining power below the threshold automatically returns to base for charging. Faulty equipment is replaced by backup equipment. Through the division of labor and collaboration among multiple equipment, the monitoring operation is ensured to be continuous.

Images