A cloud edge terminal cooperative global electromagnetic spectrum intelligent perception and dynamic scheduling system and method

Abstract

The application discloses a kind of cloud edge-end coordination global electromagnetic spectrum intelligent perception and dynamic scheduling system and method, belong to electromagnetic spectrum intelligent perception and resource scheduling technical field.The application is based on the localization cloud native architecture and constructs edge and cloud end two-level cooperative processing architecture, and edge side completes the pre-position noise reduction of multi-source heterogeneous electromagnetic signal, feature extraction and compression processing, supports offline autonomous operation, effectively reduces data transmission delay and bandwidth overhead;Cloud end is iterated and updated by fingerprint library, realizes the fusion iteration of global electromagnetic feature, the intelligent identification of radiation source, the analysis of spectrum behavior law and the deduction of tactical intention, and dynamically reconstructs perception resource according to real-time electromagnetic situation and node operating state.The application solves the technical problems of high delay, insufficient intelligent cognitive ability, rigid resource scheduling, weak invulnerability and non-domestic of existing spectrum processing scheme, can stably realize millisecond-level global spectrum situation awareness in complex strong countermeasure electromagnetic environment, with the characteristics of autonomous iteration, dynamic adaptation, high reliability and autonomous control, suitable for various global electromagnetic environment monitoring and spectrum dynamic scheduling scene.

Description

Technical Field

[0001] This invention belongs to the fields of electromagnetic spectrum sensing, distributed edge computing, intelligent signal interpretation and dynamic resource scheduling technology, specifically involving a cloud-edge-device collaborative full-domain electromagnetic spectrum intelligent sensing and dynamic scheduling system and method based on a domestically developed cloud-native architecture. Background Technology

[0002] In complex electromagnetic environments, the types of spatial electromagnetic signals are diverse, the number of radiation sources is large, and the signal states change drastically over time, placing high demands on the real-time performance, completeness, intelligence, and system resilience of spectrum sensing. Existing electromagnetic spectrum monitoring and processing technologies have significant shortcomings.

[0003] Existing technologies mostly employ a centralized data processing model, where all raw electromagnetic data acquired by front-end acquisition devices is transmitted back to a central server for unified processing. This results in large data transmission volumes, long processing links, and high overall response latency, failing to meet the real-time requirements for millisecond-level situational awareness output. In scenarios involving electromagnetic countermeasures and interference, or unstable communication links, centralized processing is prone to data interruptions and area sensing failures, leading to poor system reliability.

[0004] Meanwhile, traditional spectrum analysis methods mostly use fixed feature template matching, which can only identify known typical signals, cannot accumulate and learn the features of unknown electromagnetic signals, lack the ability to continuously iterate and update electromagnetic features, have difficulty accurately distinguishing various heterogeneous radiation sources in complex scenarios, cannot effectively analyze signal emission patterns and behavioral characteristics, and have limited situational awareness capabilities.

[0005] In addition, the computing resources, acquisition parameters, and sensing strategies of existing spectrum monitoring systems are mostly fixed configurations, which cannot dynamically adjust the resource allocation method according to real-time electromagnetic environment changes. The matching degree between sensing resources and actual electromagnetic situation is low, the resource utilization rate is uneven, and it is difficult to adapt to complex electromagnetic scenarios with high dynamics and strong changes.

[0006] Furthermore, most existing processing systems rely on foreign underlying architecture components and lack domestically developed, self-controllable capabilities, posing security risks in critical application scenarios. In short, there is currently a lack of an electromagnetic spectrum collaborative processing solution that can achieve edge-front processing, full-domain feature iteration, intelligent situational awareness, dynamic resource scheduling, and is compatible with domestically developed architectures. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of existing technologies, such as high latency in spectrum sensing, weak cognitive ability, rigid resource scheduling, poor environmental adaptability, and lack of autonomous control, and to provide a cloud-edge-device collaborative global electromagnetic spectrum intelligent sensing and dynamic scheduling system and method to achieve high-precision, low-latency, intelligent, and highly reliable spectrum situation sensing and dynamic resource allocation in complex electromagnetic environments.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: The system of this invention is built on a domestically produced cloud-native architecture and adopts a two-level collaborative processing architecture of edge and cloud. It does not rely on large platform components. Through the combination of lightweight edge processing and cloud intelligent scheduling, it can complete the collaborative perception and dynamic control of electromagnetic signals across the entire domain. This invention performs electromagnetic signal preprocessing at the edge, denoising, extracting features, and compressing the raw data, uploading only valid feature vectors, significantly reducing data transmission volume and processing latency. It also possesses offline independent operation capabilities, ensuring uninterrupted local spectrum monitoring in communication anomaly scenarios, and improving the system's anti-interference and resilience. This invention achieves the fusion, iteration, intelligent interpretation, and dynamic scheduling of global electromagnetic features in the cloud. By continuously aggregating edge feature data, it forms a global electromagnetic fingerprint data system that can be updated online, solving the problem of poor adaptability in traditional template matching. Relying on deep learning models to perform in-depth analysis of electromagnetic signals, it enables radiation source identification, behavioral pattern analysis, and tactical intent deduction, improving the level of spectrum situation awareness. By combining node load and dynamic allocation of computing power and acquisition resources based on the electromagnetic situation, it achieves adaptive reconstruction of sensing resources. This invention effectively shortens the latency of spectrum status updates and improves the adaptability and intelligence level of complex electromagnetic environments through a closed-loop process of edge preprocessing, cloud-based iterative cognition, and dynamic scheduling and control.

[0009] Compared with the prior art, the present invention has the following advantages: 1. This invention is based on a domestically developed cloud-native architecture, with an independently controllable underlying architecture, avoiding security risks brought by foreign components, and adapting to the application needs of key fields. 2. This invention adopts an edge lightweight processing mechanism, which greatly reduces data transmission pressure and redundant calculation, effectively reduces situation update latency, and can achieve millisecond-level spectrum situation output. 3. This invention has the capability of incremental iteration of electromagnetic features, can autonomously accumulate unknown signal features, continuously optimize recognition capabilities, and adapt to dynamically changing complex electromagnetic environments. 4. This invention enables cognitive-level spectrum analysis, which can deeply explore the spectrum behavior characteristics and tactical intentions of radiation sources, breaking through the technical limitations of traditional simple signal detection. 5. This invention supports dynamic reconstruction of sensing resources and offline edge autonomy, and its environmental adaptability, operational stability and resource utilization are significantly better than the traditional fixed configuration mode. Attached Figure Description Figure 1 This is a schematic diagram illustrating the overall system architecture and data interaction of the present invention; Figure 2 This is a schematic diagram of the whole-domain electromagnetic spectrum intelligent sensing and dynamic scheduling method of the present invention; Figure 3 This is a schematic diagram of the dynamic reconfiguration and scheduling process of spectrum resources according to the present invention. Detailed Implementation

[0010] The present invention will be further described in detail below with reference to specific embodiments. The system in this embodiment is deployed based on a domestic operating system and a domestic cloud-native container environment, and is divided into two parts: an edge collaborative processing module and a cloud electromagnetic cognition and scheduling module. The edge collaborative processing module is deployed on each edge computing node, connecting in real time to multiple heterogeneous electromagnetic signal acquisition devices to continuously receive raw spatial electromagnetic sampling data. The edge module filters and reduces noise from the raw data, removes invalid data, and extracts multi-dimensional information such as signal frequency, amplitude, phase, modulation parameters, and pulse characteristics, compressing it to generate standardized electromagnetic feature vectors. Edge nodes locally cache regional feature data, continuously pushing incremental feature data to the cloud when the network is normal. When electromagnetic interference causes communication link fluctuations or interruptions, the edge module automatically activates offline working logic, continuously generating local spectrum status based on locally cached data to complete regional signal monitoring and anomaly identification. After the link is restored, it automatically fills in the missing data, achieving seamless continuity. The cloud-based fingerprint database iterative update unit receives incremental feature vectors uploaded from each edge node in real time, fuses and registers multi-source, heterogeneous, and time-dispersed electromagnetic feature data, and integrates them to form full-domain electromagnetic fingerprint data. It compares and verifies the features of registered radiation sources, correcting deviation parameters; it automatically classifies and incrementally adds newly appearing unknown electromagnetic signal features to the database, completing iterative optimization of fingerprint data without interrupting system operations. The cloud-based intelligent electromagnetic signal interpretation unit loads a trained deep learning inference model and, based on global electromagnetic fingerprint data, classifies and identifies real-time aggregated electromagnetic features to accurately determine the types of various radiation source equipment. Combining this with continuous temporal spectrum data, it analyzes behavioral patterns such as signal transmission cycles, occupied frequency bands, and activity periods. By inferring the operational status and tactical intentions of radiation sources based on scene characteristics, it outputs structured, globally unified, cognitive-level spectrum situational awareness results. The cloud-based spectrum resource dynamic scheduling unit collects real-time data on the computing power utilization rate, communication bandwidth margin, regional electromagnetic signal density, and threat level of each edge node, dynamically balancing the allocation of cloud inference computing power and edge processing computing power. For areas with active electromagnetic signals, complex interference, and high threat levels, it issues instructions to improve the sampling accuracy and reporting frequency; for areas with stable signals, it reduces the investment in acquisition resources, achieving dynamic reconstruction and on-demand allocation of sensing resources. The situation output unit aggregates global spectrum situation data, identifies abnormal electromagnetic signals and illegal radiation activities in real time, generates alarm information and control instructions and sends them to edge nodes, forming a complete closed-loop perception and scheduling process, and stably achieves millisecond-level global spectrum situation update output.

Claims

1. A cloud-edge-device collaborative global electromagnetic spectrum intelligent sensing and dynamic scheduling system, characterized in that, Deployed based on a domestically developed cloud-native architecture, including an edge collaborative processing module and a cloud-based electromagnetic cognition and scheduling module; The edge collaborative processing module is deployed on the edge computing node to receive multiple heterogeneous electromagnetic signal data in real time, perform signal noise reduction, feature extraction and feature compression, and generate standardized electromagnetic feature vectors. The edge collaborative processing module locally caches the electromagnetic feature data, transmits the feature vectors upward when the communication link is normal, independently outputs the local spectrum situation when the link is abnormal, and automatically synchronizes incremental data after the link is restored. The cloud-based electromagnetic cognition and scheduling module is deployed on a domestically produced cloud-native container cluster, including a fingerprint database iterative update unit, an electromagnetic signal intelligent interpretation unit, a spectrum resource dynamic scheduling unit, and a situation output unit; The fingerprint database iterative update unit is used to aggregate the electromagnetic feature vectors uploaded by each edge node, integrate them to form a full-domain electromagnetic fingerprint dataset containing multi-dimensional features in the time domain, frequency domain, spatial domain, and modulation domain, perform matching and verification of known radiation source features, and perform incremental input and iterative update of unknown electromagnetic signal features. The electromagnetic signal intelligent interpretation unit is used to analyze the global electromagnetic feature data based on the deep learning inference model, complete the identification of radiation source type, the mining of spectrum time sequence activity patterns, and infer the target tactical intentions based on spectrum behavior characteristics, and output a global unified spectrum cognitive situation. The spectrum resource dynamic scheduling unit is used to monitor the computing load, communication status and regional electromagnetic signal density of each edge node in real time, and dynamically adjust the edge signal acquisition parameters and computing power allocation strategy according to the global spectrum situation to complete the dynamic reconstruction of electromagnetic sensing resources. The situation output unit is used to summarize global spectrum situation information, generate abnormal signal alarm information and output control and scheduling instructions. The system achieves millisecond-level global spectrum situational awareness in complex electromagnetic environments through the coordinated efforts of edge lightweight preprocessing, cloud-based intelligent cognition, and dynamic scheduling.

2. The cloud-edge-device collaborative full-domain electromagnetic spectrum intelligent sensing and dynamic scheduling system according to claim 1, characterized in that, The edge collaborative processing module only transmits the compressed electromagnetic feature vector to the cloud, without transmitting the original sampling data, thereby reducing the transmission bandwidth usage and backend processing pressure.

3. The cloud-edge-device collaborative full-domain electromagnetic spectrum intelligent sensing and dynamic scheduling system according to claim 1, characterized in that, The fingerprint database iterative update unit supports online incremental updates, completing the input of unknown signal features and optimization of feature parameters while the system is running continuously, without the need for shutdown and restart.

4. The cloud-edge-device collaborative full-domain electromagnetic spectrum intelligent sensing and dynamic scheduling system according to claim 1, characterized in that, The spectrum resource dynamic scheduling unit increases the sampling frequency and computing power ratio in areas with high electromagnetic signal density, and reduces the acquisition resource overhead in low-activity areas, thereby achieving adaptive allocation of sensing resources.

5. A cloud-edge-device collaborative method for intelligent sensing and dynamic scheduling of the entire electromagnetic spectrum, characterized in that, The system applied to any one of claims 1 to 4 includes the following steps: S1. The edge collaborative processing module receives multi-source heterogeneous electromagnetic signals in real time, extracts multi-dimensional electromagnetic features and compresses them to generate feature vectors, and caches them to form local electromagnetic feature data. S2. The edge collaborative processing module executes corresponding processing logic according to the communication link status. When the link is normal, it uploads incremental feature vectors to the cloud. When the link is interrupted, it generates local spectrum situation offline. After the link is restored, it fills in the synchronization data. S3, the cloud fingerprint database iterative update unit gathers global edge feature data, completes multi-source feature fusion, comparison and verification and incremental update, and constructs dynamic global electromagnetic fingerprint data; S4, the cloud-based electromagnetic signal intelligent interpretation unit, combines global electromagnetic fingerprint data and uses a deep learning model to complete electromagnetic signal classification, radiation source identification, spectrum behavior pattern analysis and tactical intent simulation, generating a global cognitive spectrum situation. S5, the cloud-based spectrum resource dynamic scheduling unit dynamically allocates computing resources and adjusts edge acquisition parameters based on the global spectrum situation, node load and communication status, to achieve dynamic reconstruction of spectrum sensing resources; S6. Output full-domain spectrum status and abnormal alarm information through the status output unit, issue scheduling and control commands, and complete closed-loop iterative perception.

6. 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 steps of the method described in claim 5.

Images