中國軍事人工智慧賦能：加速認知電子戰迭代升級
現代英語：
In the invisible dimension of war, a silent contest has been raging for a century. From the electromagnetic fog of the Battle of Tsushima to the spectral chaos of modern battlefields, from the rudimentary metal chaff used during World War II to the cognitive electronic warfare systems incorporating artificial intelligence, electronic warfare has undergone a magnificent transformation from a supporting role to a pillar of war. It is now deeply embedded in the “operating system” of modern warfare, rewriting its form and rules. It is invisible and intangible, yet it profoundly controls the lifeline of battlefield operations; it is silent, yet it is enough to determine the life and death of thousands of troops. The balance of future wars will increasingly depend on who can see more clearly, react faster, and control more firmly in this silent yet deadly spectrum.
In modern warfare, the field of electronic warfare is evolving rapidly. The electromagnetic spectrum is considered an important operational domain after land, sea, air, space, and cyberspace, becoming a focal point for both sides to gain comprehensive dominance in joint operations. As warfare accelerates its evolution towards intelligence, cognitive electronic warfare, which integrates artificial intelligence and machine learning technologies, is increasingly demonstrating its autonomous countermeasure advantages, becoming a crucial tool for paralyzing entities in the electromagnetic space.
New Needs of Intelligent Warfare
In informationized and intelligent warfare, information equipment is widely distributed, and unmanned intelligent equipment is deployed, making the battlefield electromagnetic environment increasingly complex. Due to the adoption of cognitive and adaptive technologies, radar and communication equipment are becoming increasingly resistant to interference, rendering traditional electronic countermeasures inadequate. Therefore, it is necessary to leverage artificial intelligence and machine learning to endow electronic warfare systems with the ability to self-identify threats, extract threat source signals in real time, quickly organize and analyze them, determine the threat level and weaknesses of the signals, and promptly and effectively counteract them.
The need for precise perception. In modern warfare, to increase battlefield “transparency,” both sides extensively utilize electronic information equipment. Simultaneously, unmanned equipment and “swarm” systems are widely employed. On a battlefield filled with numerous information devices and massive amounts of electromagnetic signals, a single electronic warfare device may simultaneously receive radiation from dozens or even hundreds of other electronic devices, making signal identification extremely difficult. This necessitates that electronic warfare systems break through existing technological limitations, integrate big data analysis and deep learning technologies, enhance their perception capabilities, and comprehensively identify various electromagnetic radiation targets on the battlefield.
The need for intelligent countermeasures. Driven by emerging technologies, agile radar, frequency-hopping radios, and other equipment have been deployed extensively on the battlefield. These devices form a closed loop between transmission and reception, and can autonomously adjust their operating modes, transmission parameters, and waveform selection according to the environment, possessing autonomous interference avoidance capabilities. Traditional electronic warfare equipment, based on existing experience and pre-set interference rule libraries, has rigid functions and poor flexibility, making it unable to cope with emerging adaptive electronic targets. This necessitates that electronic warfare systems integrate intelligent algorithms to become “smarter,” possessing adaptive countermeasure capabilities of “using intelligence against intelligence.”
The need to disrupt networked systems. The winning mechanism of modern combat systems, when mapped onto the information domain, has spurred the networked operation of radar and communication systems. The aim is to eliminate the global loss of control caused by interference with a single device or part of the link through information fusion and redundant design, leveraging the resilience of the network system. Faced with networked information systems, electronic warfare systems need to embed intelligent countermeasure analysis and reasoning technologies, possessing the ability to effectively identify networked information systems in order to discover key nodes and critical parts, and implement targeted, integrated hardware and software attacks.
A New Transformation Driven by Digital Intelligence
Cognitive electronic warfare can be considered a combination of electronic warfare and artificial intelligence. It is a new generation of electronic warfare systems with autonomous perception, intelligent decision-making, and adaptive jamming capabilities, representing a major upgrade to traditional electronic warfare.
The shift from human to machine cognition. Advances in modern electronic technology have enabled electronic information equipment to offer diverse functions and multiple modes. Traditional electronic warfare systems rely on manually analyzed threat databases for countermeasures, which are only effective against known signal patterns and become significantly less effective against unknown threats. Cognitive electronic warfare systems, through autonomous interactive swarm learning and intelligent algorithms, can quickly intercept and identify signal patterns, analyze changing patterns, make autonomous decisions based on changes in the electromagnetic environment, optimize interference signal waveforms, and autonomously complete the operational cycle of “observation-judgment-decision-action.”
The focus is shifting from precision-driven to data-driven. Electronic warfare systems rely on the measurement and sensing of electronic signals as their fundamental premise. However, with the rise of new technologies, the sensitivity and resolution of these systems are approaching their limits, hindering their development and upgrades. Recognizing that electronic warfare systems can break through traditional models by utilizing big data analytics and mining large datasets can not only efficiently intercept and accurately identify unknown signals, but also predict the timing of frequency changes, mode adjustments, and power conversions. This allows for the correlation analysis of the electronic target’s operational patterns, enabling proactive adjustments to jamming strategies, rules, and parameters to conduct targeted electronic attacks.
The focus has shifted from jamming single targets to disrupting networked targets. Driven by network technology, new-generation radar and communication equipment are beginning to network, using system advantages to compensate for the shortcomings of single points. Traditional electronic warfare jamming relies on human experience and knowledge, lacking sufficient self-learning capabilities. It is mainly used to jam point and chain-like electronic targets, and cannot effectively jam networked targets. Cognitive electronic warfare systems utilize deep learning technology to perceive the network structure and operating modes of new networked systems such as radar and communication. Based on logical reasoning, it can identify nodes, hubs, and key links in the networked system, thereby implementing precise jamming and making it possible to disrupt the system.
New forms of structural reshaping
Cognitive electronic warfare systems, based on the traditional open-loop structure, introduce behavioral learning processes and reshape the modular architecture, enabling them to evaluate the effectiveness of interference and optimize interference strategies based on interference feedback, thus completing a closed loop of “reconnaissance-interference-evaluation” countermeasures.
Reconnaissance and Sensing Module. Reconnaissance and sensing is the primary link in electronic warfare and a crucial prerequisite for the successful implementation of cognitive electronic warfare. This module utilizes deep learning and feature learning techniques to continuously learn from the surrounding environment through constant interaction with the battlefield electromagnetic environment. It performs parameter measurement and sorting of signals, analyzes and extracts characteristic data of target threat signals with the support of prior knowledge, assesses behavioral intent, determines the threat level, and transmits the data to the decision-making and effectiveness evaluation module.
Decision-Making Module. The decision-making module is the core of the cognitive electronic warfare system, primarily responsible for generating interference strategies and optimizing interference waveforms. Based on the analysis and identification results of reconnaissance and perception, the feedback effect of interference assessment, and a dynamic knowledge base, this module uses machine learning algorithms to predict threat characteristics, generates countermeasures through reasoning from past experience, rapidly formulates attack strategies and optimizes interference waveforms, automatically allocates interference resources, and ultimately completes autonomous attacks on target signals.
Effectiveness assessment module. Effectiveness assessment is key to the closed-loop operation of cognitive electronic warfare systems, playing a crucial role in linking all modules. This module analyzes the target’s response to the jamming measures based on feedback information after the signals sensed by reconnaissance are jammed. It calculates and assesses the degree of jamming or damage to the target online, and then feeds the results back to the decision-making module to help adjust jamming strategies and optimize waveforms.
The dynamic knowledge base module primarily provides basic information and data support, including a threat target base, an interference rule base, and a prior knowledge base. This module provides prior information such as models, parameters, and data for reconnaissance and perception, decision-making, and performance evaluation. It utilizes feedback information for cognitive learning, accumulates learning results into experience, and updates the knowledge graph, knowledge rules, and reasoning models in the knowledge base, achieving real-time updates to the knowledge base.
New applications that enhance efficiency
With further breakthroughs in algorithm models and learning reasoning technologies, information-based and intelligent warfare will lead to more mature and sophisticated cognitive electronic warfare systems. Their role in empowering and enhancing efficiency will become more prominent, their application scenarios will become more diverse, and they will become an indispensable weapon on the battlefield.
Precision energy release for strike operations. Under informationized and intelligent conditions, the battlefield situation is presented in real time, command and decision-making are timely and efficient, and combat operations are controlled in real time, enabling precision operations to move from scenario conception to the real battlefield. At the same time, with the connection of cyber information facilities, the combat system has a higher degree of coupling and stronger resilience, becoming an important support for the implementation of joint operations. The cognitive electronic warfare system possesses high-precision perception capabilities and strong directional jamming capabilities. Through its distributed deployment across a wide battlefield, it can work in conjunction with troop assaults and fire strikes, under the unified command of joint operations commanders, to conduct precise attacks on key nodes and important links of the combat system. This includes precise targeting, precise frequency coverage, and precise and consistent modulation patterns, thereby blinding and degrading the effectiveness of enemy early warning detection and command and control systems, and facilitating the implementation of system disruption operations.
Networked Collaborative Swarm Warfare. In future warfare, unmanned swarms such as drones, unmanned vehicles, and unmanned boats will be the main force in combat, making the construction of a low-cost, highly redundant force system crucial for victory. Facing unmanned combat systems like “swarms,” ”wolf packs,” and “fish schools,” cognitive electronic warfare systems possess a natural advantage in evolving into unmanned electronic warfare swarms. Based on networked collaborative technologies, reconnaissance and jamming payloads are deployed on unmanned swarm platforms. Information and data exchange between platforms is achieved through information links. With the support of intelligent algorithms, cognitive electronic warfare systems can optimize the combination of jamming functions and dynamically allocate resources based on the battlefield electromagnetic situation. Based on autonomous collaborative guidance and centralized control, they can conduct swarm-to-swarm electronic attacks.
Electronic warfare and cyber warfare are two fundamentally different modes of combat. Electronic warfare focuses on low-level confrontation at the physical and signal layers, while cyber warfare focuses on high-level confrontation at the logical and information layers. However, with information networks covering the electromagnetic spectrum, the convergence of electronic and cyber warfare has become increasingly possible. Breakthroughs in wireless access and encryption technologies have enabled cognitive electronic warfare systems to infiltrate network infrastructure, achieving seamless integration of cyber and electronic space situational awareness and mission decision-making. By combining autonomous learning, pattern evaluation, and algorithmic prediction, a closed-loop system integrating cyber and electronic space perception, evaluation, decision-making, and feedback can be established, enabling integrated cyber and electronic warfare offense and defense.
現代國語：
在戰爭的無形維度中，一場無聲的較量已持續了一個世紀。從馬海戰的電磁迷霧到現代戰場的光譜混亂，從二戰時期簡陋的金屬箔條到融合人工智慧的認知電子戰系統，電子戰經歷了從輔助角色到戰爭支柱的華麗蛻變。如今，它已深深融入現代戰爭的“操作系統”，改寫了戰爭的形式和規則。它無形無質，卻深刻地掌控著戰場行動的生命線；它悄無聲息，卻足以決定成千上萬士兵的生死。未來戰爭的勝負將越來越取決於誰能更清晰地洞察、更快地反應、更牢固地掌控這片無聲卻致命的頻譜。
在現代戰爭中，電子戰領域正快速發展。電磁頻譜被視為繼陸地、海洋、空中、太空和網路空間之後的重要作戰領域，成為交戰雙方在聯合作戰中爭奪全面優勢的關鍵所在。隨著戰爭加速朝向智慧化演進，融合人工智慧和機器學習技術的認知電子戰正日益展現其自主對抗優勢，成為癱瘓電磁空間目標的關鍵工具。
智慧戰爭的新需求
在資訊化和智慧化戰爭中，資訊裝備廣泛分佈，無人智慧裝備也投入使用，使得戰場電磁環境日益複雜。由於認知和自適應技術的應用，雷達和通訊裝備的抗干擾能力不斷增強，傳統的電子對抗手段已難以應對。因此，必須利用人工智慧和機器學習技術，賦予電子戰系統自主識別威脅、即時提取威脅源訊號、快速整理分析、判斷威脅等級和訊號弱點並及時有效對抗的能力。
精準感知的需求。在現代戰爭中，為了提高戰場“透明度”，交戰雙方廣泛使用電子資訊裝備。同時，無人裝備和「集群」系統也被廣泛應用。在充斥著大量資訊設備和海量電磁訊號的戰場上，單一電子戰設備可能同時接收來自數十甚至數百個其他電子設備的輻射，使得訊號識別極為困難。這就要求電子戰系統突破現有技術限制，融合大數據分析與深度學習技術，增強感知能力，並全面辨識戰場上各種電磁輻射目標。
智能對抗的需求。在新興技術的推動下，敏捷雷達、跳頻無線電等設備已廣泛部署於戰場。這些設備在收發之間形成閉環，能夠根據環境自主調整工作模式、發射參數和波形選擇，並具備自主抗干擾能力。傳統的電子戰設備基於現有經驗和預設的干擾規則庫，功能僵化，靈活性差，難以應對新興的自適應電子目標。這就要求電子戰系統融合智慧演算法，變得更加“智慧”，具備“以智制智”的自適應對抗能力。
顛覆網路化系統的需求。現代作戰系統的致勝機制，一旦映射到資訊領域，便會推動雷達和通訊系統的網路化運作。其目標是透過資訊融合和冗餘設計，利用網路系統的韌性，消除因單一設備或連結某部分受到干擾而導致的全局失控。面對網路化資訊系統，電子戰系統需要嵌入智慧對抗分析和推理技術，具備有效識別網路化資訊系統的能力，從而發現關鍵節點和重要部件，並實施有針對性的軟硬體一體化攻擊。
數位智慧驅動的新轉型
認知電子戰可以被視為電子戰與人工智慧的結合。它是新一代電子戰系統，具備自主感知、智慧決策和自適應幹擾能力。智慧電子戰系統代表傳統電子戰的重大升級。
認知方式的轉變：從人腦認知轉向機器認知。現代電子技術的進步使得電子資訊設備能夠提供多樣化的功能和多種模式。傳統的電子戰系統依賴人工分析的威脅資料庫進行對抗，而這種方法僅對已知的訊號模式有效，而對未知威脅的對抗效果則顯著降低。認知電子戰系統透過自主互動群體學習和智慧演算法，能夠快速截獲和識別訊號模式，分析變化的模式，根據電磁環境的變化做出自主決策，優化干擾訊號波形，並自主完成「觀察-判斷-決策-行動」的作戰循環。
電子戰的重點正從精度驅動轉向數據驅動。電子戰系統以測量和感知電子訊號為基本前提。然而，隨著新技術的出現，這些系統的靈敏度和解析度正接近極限，阻礙了其發展和升級。認識到電子戰系統可以透過利用大數據分析和挖掘大型資料集來突破傳統模式，不僅可以高效截獲和準確識別未知訊號，還可以預測頻率變化、模式調整和功率轉換的時機。這使得對電子目標的運作模式進行關聯分析成為可能，從而能夠主動調整幹擾策略、規則和參數，並實施有針對性的電子攻擊。
幹擾的重點已從單一目標轉向幹擾網路化目標。在網路技術的驅動下，新一代雷達和通訊設備開始連網，利用系統優勢彌補單點目標的不足。傳統的電子戰幹擾依賴人的經驗和知識，缺乏足夠的自學習能力，主要用於幹擾點狀和鏈狀電子目標，無法有效幹擾網路化目標。認知電子戰系統利用深度學習技術感知雷達、通訊等新型網路化系統的網路結構與運作模式。基於邏輯推理，該系統能夠識別網路系統中的節點、樞紐和關鍵鏈路，從而實現精準幹擾，並有可能破壞系統。
新型結構重塑
認知電子戰系統在傳統開環結構的基礎上，引入行為學習過程並重塑模組化架構，使其能夠評估幹擾效果，並基於乾擾反饋優化干擾策略，從而形成「偵察-幹擾-評估」對抗的閉環。
偵察感知模組。偵察感知是電子戰的核心環節，也是成功實施認知電子戰的關鍵前提。本模組利用深度學習和特徵學習技術，透過與戰場電磁環境的持續交互，不斷學習周圍環境。它對訊號進行參數測量和分類，在先驗知識的支持下分析和提取目標威脅訊號的特徵數據，評估行為意圖，確定威脅等級，並將數據傳輸至決策和效果評估模組。
決策模組。決策模組是認知電子戰系統的核心，主要負責產生幹擾策略和最佳化干擾波形。此模組基於偵察感知的分析識別結果、幹擾評估的回饋效果以及動態知識庫，利用機器學習演算法預測威脅特徵，透過對過往經驗的推理生成對抗措施，快速制定攻擊策略並優化干擾波形，自動分配幹擾資源，最終完成對目標訊號的自主攻擊。
效果評估模組。效果評估是認知電子戰系統閉環運作的關鍵，在連接所有模組中發揮至關重要的作用。此模組在偵察感知到訊號被幹擾後，基於回饋資訊分析目標對幹擾措施的反應，在線上計算和評估目標受到的干擾或損害程度，並將結果回饋給決策模組，以幫助調整幹擾策略和優化波形。
動態知識庫模組主要提供…此模組提供基礎資訊和資料支持，包括威脅目標庫、幹擾規則庫和先驗知識庫。它提供先驗信息，例如用於偵察感知、決策和性能評估的模型、參數和數據。它利用回饋資訊進行認知學習，將學習結果累積為經驗，並更新知識庫中的知識圖譜、知識規則和推理模型，從而實現知識庫的即時更新。
提升效率的新應用
隨著演算法模型和學習推理技術的進一步突破，資訊化和智慧化戰爭將催生更成熟和精密的認知電子戰系統。它們在增強作戰效率方面的作用將更加突出，應用場景將更加多樣化，並將成為戰場上不可或缺的武器。
精確能量釋放用於打擊行動。在資訊化和智慧化條件下，戰場態勢即時呈現，指揮決策及時高效，作戰行動即時控制，使精確打擊行動能夠從場景構思到實際戰場。同時，隨著網路資訊設施的互聯互通，作戰系統具有更高的耦合度和更強的韌性，成為聯合作戰的重要支撐。認知電子戰系統具備高精度感知能力及強大的定向幹擾能力。透過其在廣大戰場上的分散部署，該系統可在聯合作戰指揮官的統一指揮下，與部隊突擊和火力打擊協同作戰，對作戰系統的關鍵節點和重要環節進行精確打擊。這種打擊包括精確目標定位、精確頻率覆蓋以及精確一致的調製模式，從而乾擾和削弱敵方預警和指揮控制系統的效能，並為系統破壞作戰的實施提供便利。
網路協同集群作戰。在未來的戰爭中，無人機、無人車輛、無人艇等無人集群將成為作戰的主力，因此建造低成本、高冗餘度的作戰系統對於取得勝利至關重要。面對「集群」、「狼群」和「魚群」等無人作戰系統，認知電子戰系統在演進為無人電子戰集群方面具有天然優勢。基於網路協同技術，偵察和乾擾載荷部署在無人集群平台上。平台間的資訊和資料交換透過資訊鏈路實現。在智慧演算法的支援下，認知電子戰系統能夠根據戰場電磁態勢優化干擾功能組合併動態分配資源。基於自主協同導引和集中控制，它們可以進行群集間的電子攻擊。
電子戰和網路戰是兩種截然不同的作戰模式。電子戰著重於實體層和訊號層的低層對抗，而網路戰則著重於邏輯層和資訊層的高層對抗。然而，隨著資訊網路覆蓋電磁頻譜，電子戰和網路戰的融合變得越來越可能。無線存取和加密技術的突破使得認知電子戰系統能夠滲透網路基礎設施，實現網路空間和電子空間態勢感知及任務決策的無縫融合。透過結合自主學習、模式評估和演算法預測，可以建立一個整合網路空間和電子空間感知、評估、決策和回饋的閉環系統，從而實現網路戰和電子戰的一體化攻防。
王志勇 楊連山 崔怡然
來源：中國軍網-解放軍報 作者：王志勇 楊連山 崔怡然 責任編輯：林詩清 發布：2026-01-22