Adaptive interference suppression carrier communication system based on complex electromagnetic environment

Abstract

The invention provides an adaptive interference suppression carrier communication system based on a complex electromagnetic environment, relates to the technical field of communication systems, and solves the problems that an existing communication system is poor in adaptability, weak in anti-interference capability and lack of collaboration and safety in the complex electromagnetic environment. The system comprises an environment perception module used for deeply perceiving an environment and identifying known and unknown interferences by using AI; the cognitive decision engine is used for generating a multi-domain control instruction by utilizing AI according to environment, internal, cooperation and safety states; the self-adaptive transceiving control module is used for executing cross-multi-domain parameter self-adaptive adjustment; the distributed collaborative module is used for realizing information sharing among nodes, federated learning and collaborative resource management; and the endogenous security enhancement module is used for executing active security mechanisms such as dynamic heterogeneous redundancy, mimicry defense and physical layer authentication. The invention further discloses a corresponding communication method. According to the method, the robustness, adaptability, efficiency and safety of a communication system in a complex electromagnetic environment can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of communication systems, and specifically to an adaptive interference suppression carrier communication system for complex electromagnetic environments. Background Art

[0002] With the rapid development and wide application of wireless communication technologies, electromagnetic spectrum resources are becoming increasingly crowded, and the electromagnetic environment faced by wireless communication systems is also becoming increasingly complex. A complex electromagnetic environment usually contains various types of interference signals with time-varying power, uncertain directions, and even signals with spoofing or intelligent countermeasure properties. At the same time, it is also superimposed with channel effects such as multipath fading and Doppler frequency shift.

[0003] Existing communication anti-interference technologies, such as simple filtering, spread spectrum, frequency hopping, or beamforming with fixed parameters, are often designed for specific and known interference patterns. When facing unknown, dynamic, and diverse complex electromagnetic environments, they have the following limitations:

[0004] Lack of cooperation ability: Most systems are optimized for single points and lack the cooperative anti-interference ability between network nodes, making it difficult to cope with distributed interference.

[0005] These limitations lead to a significant decline in the reliability, availability, and data transmission efficiency of communication systems in complex electromagnetic environments, making it difficult to meet the stringent requirements for communication resilience in scenarios such as emergency communication and critical industrial control. Therefore, there is an urgent need for an adaptive interference suppression communication system that can intelligently sense the environment, make autonomous decisions, work cooperatively, and have built-in security capabilities. Summary of the Invention

[0006] Technical Problems to be Solved

[0007] Aiming at the deficiencies of the prior art, the present invention provides an adaptive interference suppression carrier communication system for complex electromagnetic environments, which solves the technical problems of poor adaptability, limited anti-interference ability, lack of cooperation and built-in security mechanisms in communication systems in complex electromagnetic environments in the prior art.

[0008] Technical Solutions

[0009] To achieve the above objectives, the present invention is realized through the following technical solutions: An adaptive interference suppression carrier communication system for complex electromagnetic environments, and the adaptive interference suppression carrier communication system includes the following components:

[0010] An environmental perception module, which is used for: monitoring electromagnetic signals within a preset frequency band, extracting multi-dimensional features of the electromagnetic signals, and generating environmental situation information including interference information and channel state information based on the multi-dimensional features; identifying the types of detected interference signals based on the multi-dimensional features, and using a pre-trained classification model or anomaly detection algorithm to identify unknown types of interference signals; and including the interference type identification result or unknown interference indication in the environmental situation information;

[0011] The multi-dimensional features include one or more of spectral features, time-domain features, and spatial-domain features;

[0012] A cognitive decision-making engine, which is used for: receiving environmental situation information and the internal state information of the adaptive interference suppression carrier communication system; using a pre-trained artificial intelligence model to generate control instructions covering multiple communication parameter domains according to the environmental situation information and internal state information;

[0013] The multiple communication parameter domains include two or more of the frequency domain, time domain, spatial domain, power domain, waveform domain, coding domain, and modulation domain;

[0014] An adaptive transceiver control module, which is used for: receiving the control instructions; configuring the communication parameters of the physical layer or media access control layer of the system according to the control instructions for performing adaptive adjustment across the multiple communication parameter domains;

[0015] A distributed cooperation module, which is used for: performing information interaction with other communication nodes in the network, and the information interaction includes sharing interference information generated by the environmental perception module and local link quality information, and providing the information obtained from other nodes to the cognitive decision-making engine as network cooperation information;

[0016] An endogenous security enhancement module, which is used for: executing at least one endogenous security mechanism according to the instructions of the cognitive decision-making engine or a preset security policy;

[0017] The endogenous security mechanism includes at least one of the following:

[0018] Dynamic heterogeneous redundancy: dynamically selecting one or a group from multiple functionally equivalent but differently implemented redundant processing modules to process communication signals;

[0019] Mimic defense: performing fast and pseudo-random parameter switching in multiple communication parameter dimensions (including frequency, time, spreading code, waveform parameters);

[0020] Physical layer authentication: Extract fingerprint information based on the channel characteristics or radio frequency characteristics of the received signal, and compare it with the fingerprint database of trusted nodes to verify the node identity.

[0021] Preferably, the spatial domain features extracted by the environmental perception module include direction-of-arrival information of signals, and the environmental situation information is configured as an interference map data structure, which includes frequency, power, direction, and type information of interference signals.

[0022] Preferably, the artificial intelligence model used by the cognitive decision-making engine is a deep reinforcement learning model; and the state space input of the deep reinforcement learning model further includes network cooperation information received from other network nodes and the security state information of the system; the reward function of the deep reinforcement learning model jointly optimizes communication throughput, bit error rate, delay, power consumption, and security metrics.

[0023] Preferably, the distributed cooperation module is further configured to: participate in distributed model training based on federated learning, where multiple nodes cooperate to train at least one shared artificial intelligence model for interference recognition, channel prediction, or policy generation, and the original perception data of the nodes does not leave the local nodes during the training process, execute a distributed resource negotiation protocol, and cooperate with other nodes to perform communication resource allocation or interference suppression operations.

[0024] Preferably, the communication parameter adjustment configured by the adaptive transceiver control module (ATCM) includes at least one of the following:

[0025] Select among multiple candidate waveforms (including OFDM, FBMC, DSSS) according to the environmental situation information or service requirements, and adaptively adjust the key parameters of the selected waveform (including cyclic prefix length, subcarrier spacing, filter parameters);

[0026] Select among multiple channel coding schemes (including LDPC, PolarCode, FountainCode) according to the environmental situation information (interference type and channel quality), and adaptively adjust the coding rate and modulation method;

[0027] Control the multi-antenna system to perform adaptive beamforming or spatial nulling according to the environmental situation information (DOA of interference signals).

[0028] Preferably, the adaptive interference suppression carrier communication method includes the following steps:

[0029] Sp1. Use the Environment Sensing Module (ESM) to monitor electromagnetic signals within a preset frequency band, extract multi-dimensional features of the electromagnetic signals (including at least one or more of spectrum, time domain, and spatial domain features), and generate environment situation information containing interference information and channel state information. The environment situation information generated by the Environment Sensing Module contains direction-of-arrival information of signals and is constructed into an interference map data structure;

[0030] Sp2. Use a pre-trained classification model or anomaly detection algorithm to identify the type of interference signals or indicate unknown interference, and include the results in the environment situation information;

[0031] Sp3. Cognitive decision-making step: Use an artificially intelligent model pre-trained in the Cognitive Decision Engine (CDE) to generate control instructions covering multiple communication parameter domains (including at least two or more of frequency, time, space, power, waveform, coding, and modulation domains) based on the environment situation information and the internal state information of the system;

[0032] Sp4. Use the Adaptive Transceiver Control Module (ATCM) to configure the communication parameters of the physical layer (PHY) and the media access control layer (MAC) of the system according to the control instructions to perform adaptive adjustment across the multiple communication parameter domains.

[0033] Preferably, the artificially intelligent model used in the cognitive decision-making step is a deep reinforcement learning model, and this step further includes:

[0034] Sp3.1. Use the network cooperation information received from other network nodes and the security state information of the system as additional state inputs to the deep reinforcement learning model;

[0035] Sp3.2. Train or guide the deep reinforcement learning model by jointly optimizing the reward function for multiple communication objectives.

[0036] Preferably, the method further includes:

[0037] Distributed cooperation step: Use the Distributed Cooperation Module to share interference information and link quality information with other communication nodes in the network, and use the obtained information for the cognitive decision-making step;

[0038] Endogenous security enhancement step: Use the Endogenous Security Enhancement Module to execute at least one security mechanism among dynamic heterogeneous redundancy, mimic defense, or physical layer authentication according to the instructions generated in the cognitive decision-making step or preset policies.

[0039] Beneficial effects

[0040] The present invention provides an adaptive interference suppression carrier communication system for a complex electromagnetic environment. It has the following beneficial effects:

[0041] 1. Through intelligent environment perception and AI-driven cognitive decision-making, it can adapt to complex, dynamic, and even unknown electromagnetic environments in real-time and precisely, significantly enhancing the survivability and reliability of communication systems under strong interference. It conducts joint optimization and adaptive adjustment across multiple dimensions such as frequency, time, space, power, waveform, coding, and modulation, achieving more refined and effective interference suppression and avoidance.

[0042] 2. Through the distributed cooperation module, it realizes intelligent sharing and collaborative decision-making among nodes, enhancing the overall anti-interference ability and resource utilization efficiency of the network. It deeply integrates security mechanisms such as dynamic heterogeneous redundancy, mimic defense, and physical layer authentication into the communication system, transforming passive defense into active immunity and effectively countering intelligent and targeted attacks.

[0043] 3. The system has the ability of autonomous learning and intelligent decision-making, reducing manual intervention and enhancing operation efficiency and response speed. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is the system structure block diagram of the present invention;

[0045] Figure 2 It is the internal function block diagram of the ESM of the present invention;

[0046] Figure 3 It is the schematic diagram of the CDE working process of the present invention;

[0047] Figure 4 It is the schematic diagram of the federated learning process in the DCM of the present invention;

[0048] Figure 5 It is the schematic diagram of the mimic defense mechanism in the ESEM of the present invention;

[0049] Figure 6 It is the flowchart of the communication method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention. Specific Embodiment 1:

[0052] As Figures 1-6As shown in the figure, this embodiment provides an adaptive interference suppression carrier communication system for complex electromagnetic environments. The system aims to provide highly reliable, efficient, resilient, and secure wireless communication capabilities in complex, dynamic, and even maliciously attacked electromagnetic environments through the precise design and intelligent collaboration of its internal functional modules. The system mainly includes the following core components: an Environment Sensing Module (ESM), a Cognitive Decision Engine (CDE), an Adaptive Transceiver Control Module (ATCM), a Distributed Collaboration Module (DCM), and an Endogenous Security Enhancement Module (ESEM). These modules are interconnected through an internal high-speed data bus or dedicated interfaces and closely cooperate with the system's Radio Frequency Front End (RFFE), antenna subsystem, and protocol stack parts such as the Physical Layer (PHY) and Media Access Control Layer (MAC).

[0053] The Environment Sensing Module (ESM), referring to Figure 1 and Figure 2 , acts as the system's environmental sensor, and its core task is to accurately, comprehensively, and real-time sense and understand the electromagnetic environment in which the system is located.

[0054] Signal Monitoring and Multi-Dimensional Feature Extraction: The ESM receives broadband radio frequency signals through the antenna subsystem and RFFE, performs necessary amplification, filtering, and high-precision analog-to-digital conversion (using a 16-bit ADC with a sampling rate greater than 500 MSPS) to obtain digital intermediate frequency or baseband IQ data streams. This data stream is then sent to the internal signal processing unit (utilizing the parallel processing capabilities of an FPGA). The signal processing unit performs a series of operations:

[0055] Broadband Spectrum Analysis (calculating the PSD through FFT).

[0056] Signal Detection (identifying potential signals using a CFAR detector).

[0057] Multi-Dimensional Feature Extraction: For the detected signals, their multi-dimensional features are extracted in parallel and in detail, which includes at least one or more of spectral features, time-domain features, and spatial-domain features:

[0058] Spectral Features: Center frequency, precise bandwidth (3dB, 6dB), power (peak, average), power spectral density shape (flatness, roll-off coefficient), sideband features, number of carriers (for multi-carrier signals), etc.

[0059] Time-Domain Features: Pulse width, time of arrival (TOA), pulse repetition interval (PRI) and its jitter pattern (for pulse interference); signal envelope features, zero-crossing rate, symbol rate (for digital modulation signals), etc.

[0060] Spatial domain features: Using multi-antenna data and DOA estimation algorithm units (MUSIC, ESPRIT), calculate the direction of arrival information (DOA) of signals, including azimuth and / or elevation angle and their estimation accuracy.

[0061] Cyclostationarity features, statistical features (peak-to-average ratio, kurtosis), modulation domain features (preliminary identification), etc.

[0062] Environmental situation information generation, interference identification and unknown detection: The extracted multi-dimensional feature vectors are fed into an AI analysis unit (which can be implemented based on an SoC, an embedded GPU, or an AI co-processor).

[0063] Interference type identification: This unit uses a pre-trained classification model to identify the type of detected interference signals based on the multi-dimensional features. The model is a deep neural network, which may adopt an architecture combining CNN layers (extracting local patterns of spectrograms or time-frequency spectrograms) and LSTM / GRU layers (processing the time series characteristics or cyclic feature evolution of signals), and finally outputs the probabilities of each known interference type through a fully connected layer and Softmax. The training of the model is based on simulation and measured data containing a large number of labeled samples, and uses optimizers such as backpropagation and Adam to minimize the cross-entropy loss.

[0064] Unknown type interference identification: This unit simultaneously uses an anomaly detection algorithm to identify interference signals of unknown types. A commonly used method is to train an Autoencoder model. The model consists of an encoder and a decoder, and is trained only using known type interference samples, with the goal of minimizing the reconstruction error (MSELoss). When the feature vector of a new signal is input, if its reconstruction error after passing through the autoencoder is significantly higher than the normal level (exceeding the threshold), it is marked as an unknown interference.

[0065] Finally, the ESM integrates all analysis results (including channel state information), generates environmental situation information containing interference information and channel state information, and explicitly includes the interference type identification result or unknown interference indication in the environmental situation information.

[0066] Interference map data structure: For the convenience of CDE understanding and use, the environmental situation information is preferably constructed into an interference map data structure. This structure systematically organizes the sensing results, and can be a dynamically updated list or database, where each entry corresponds to a detected interference source or an affected frequency band / direction area, and contains information such as the frequency range of the interference signal, peak / average power, estimated direction of arrival (DOA), identified type (or marked as unknown), and a confidence score.

[0067] The ESM outputs this comprehensive, detailed, and real-time environmental situation information (including the interference map) to the CDE.

[0068] The Cognitive Decision Engine (CDE), as the intelligent decision-making center of the system, its detailed working mechanism refers to Figure 3 , Information Input and State Construction: The CDE receives the environmental situation information (from the ESM) and the internal state information of the system (from the ATCM, etc.). Importantly, it also receives network cooperation information provided by other network nodes via the DCM and system security state information provided by the ESEM. These information jointly define the state space input of the DRL model. The state information needs to be preprocessed before being input into the model.

[0069] AI-based Intelligent Decision Making:

[0070] Model Type and Implementation: The core of the CDE is to use a pre-trained artificial intelligence model for decision-making, and this model is preferably a deep reinforcement learning (DRL) model. Algorithms such as PPO (Proximal Policy Optimization) or SAC (Soft Actor-Critic) are selected.

[0071] PPO: Stabilize policy updates through a truncated surrogate objective function and estimate advantages by combining GAE.

[0072] SAC: Introduce a maximum entropy objective to encourage exploration and use Twin Q-networks and target networks to stabilize learning.

[0073] The neural network of the DRL model can include CNN, RNN (LSTM / GRU), Transformer, or a combination thereof. An architecture based on Transformer can be adopted to utilize the self-attention mechanism to process state dependencies.

[0074] Training: The DRL model is mainly obtained through offline simulation training, which requires a large amount of interaction and computing resources. The design of the reward function is crucial. Reward Function Design: The reward function of the deep reinforcement learning model is designed to jointly optimize communication throughput, bit error rate, latency, power consumption, and security metrics. Use a weighted sum R = Σw_i metric_i, and the weights w_i can be dynamically adjusted.

[0075] Multi-domain Control Instruction Generation: The Actor network of the DRL model outputs an action vector, and the CDE parses it into control instructions covering multiple communication parameter domains. These domains include at least two or more of the frequency domain, time domain, spatial domain, power domain, waveform domain, coding domain, and modulation domain. The CDE sends the control instructions to the ATCM.

[0076] Adaptive Transceiver Control Module (ATCM), Instruction Reception and Parsing: The ATCM receives the control instructions (from the CDE).

[0077] Parameter Configuration and Execution: The ATCM configures the communication parameters of the physical layer (PHY) or the media access control layer (MAC) of the system according to the control instructions for performing adaptive adjustment across multiple communication parameter domains. The specific configuration actions include at least one or more of the following:

[0078] Waveform Selection and Parameterization:

[0079] Select among multiple candidate waveforms: Select OFDM, FBMC, DSSS, etc. according to the environmental situation information or service requirements.

[0080] Adaptive adjustment of key parameters of the selected waveform: Adjust the cyclic prefix (CP) length and subcarrier spacing of OFDM; adjust the filter parameters of FBMC.

[0081] Coding and Modulation Adjustment:

[0082] Select among multiple channel coding schemes: Select LDPC, PolarCode, FountainCode, etc. according to the environmental situation information (especially the interference type and channel quality).

[0083] Adaptive adjustment of coding rate and modulation mode: Select the optimal combination from the MCS list according to the real-time link quality feedback.

[0084] Spatial Processing Control:

[0085] According to the environmental situation information (especially the DOA of the interference signal), control the multi-antenna system to perform adaptive beamforming (calculate the weight using the MMSE criterion w = αR_nn^{-1}a(θ_d)) or spatial nulling (calculate the weight using the ZF criterion w = P_nulla(θ_d)).

[0086] Distributed Collaboration Module (DCM) realizes network collaboration.

[0087] Information Interaction among Nodes: The DCM exchanges information with other communication nodes in the network. The information interaction includes at least: sharing the interference information generated by the Environmental Sensing Module (ESM) and sharing the local link quality information.

[0088] Information Supply to CDE: The DCM provides the information obtained from other nodes (possibly after fusion) to the Cognitive Decision Engine (CDE) as network collaboration information.

[0089] Advanced Collaboration Function: The DCM is also configured to perform:

[0090] Participate in distributed model training based on Federated Learning (FL): Figure 4 As shown, co-train at least one shared artificial intelligence model for interference recognition, channel prediction, or policy generation. Adopt the FedAvg process: perform local training at nodes and only upload model updates (not raw data) to the aggregator for aggregation. During the training process, the original perception data of the nodes does not leave the local nodes.

[0091] Execute the distributed resource negotiation protocol: Implement and execute the distributed resource negotiation protocol (based on signaling or optimization algorithms such as ADMM), and cooperate with other nodes to perform communication resource allocation or interference suppression operations.

[0092] The Endogenous Security Enhancement Module (ESEM) provides endogenous security capabilities.

[0093] Execute the trigger: The actions of the ESEM are initiated according to the instructions of the Cognitive Decision Engine (CDE) or the pre-set security policies of the system.

[0094] The executed endogenous security mechanisms include at least one of the following:

[0095] Dynamic Heterogeneous Redundancy (DHR): Dynamically select one or a group from multiple functionally equivalent but differently implemented redundant processing modules to process the current communication signal. The selection mechanism can be based on instructions or random policies.

[0096] Mimic Defense: Refer to Figure 5 , and the control system performs fast and pseudo-random parameter switching in multiple communication parameter dimensions (including frequency, time, spreading code, waveform parameters, etc.). The switching sequence is generated by a CSPRNG (AES-CTR), and the transceiver parties rely on the synchronization mechanism to stay consistent.

[0097] Physical Layer Authentication (PHY Authentication): Extract fingerprint information based on the channel characteristics or radio frequency characteristics of the received signal. Then, compare this fingerprint vector with the fingerprint templates of trusted nodes stored in the security database (calculate the Euclidean distance, cosine similarity, or use an SVM classifier), and only confirm the identity if the match is successful. Specific Embodiment 2:

[0099] As Figure 6 shown, the adaptive interference suppression carrier communication method includes the following steps:

[0100] Sp1. Use the Environmental Sensing Module (ESM) to monitor electromagnetic signals within a preset frequency band, extract multi-dimensional features of the electromagnetic signals (including at least one or more of spectrum, time domain, and spatial domain features), and generate environmental situation information containing interference information and channel state information. The environmental situation information generated by the environmental sensing module contains direction-of-arrival information of signals and is constructed into an interference map data structure; Environmental Sensing (Enhanced): ESM monitors signals, extracts multi-dimensional features, and generates environmental situation information. Ensure that the information contains DOA information and is constructed into an interference map data structure;

[0101] Sp2. Use a pre-trained classification model or anomaly detection algorithm to identify the type of interference signals or indicate unknown interference, and include the results in the environmental situation information. Interference Identification: The internal AI unit of ESM uses a pre-trained classification model to identify the type of interference and uses an anomaly detection algorithm to indicate unknown interference, and the results are incorporated into the environmental situation information;

[0102] Sp3. Cognitive Decision-making Step: Use the artificially intelligent model pre-trained in the Cognitive Decision-making Engine (CDE) to generate control instructions covering multiple communication parameter domains (including at least two or more of frequency, time, space, power, waveform, coding, and modulation domains) according to the environmental situation information and the internal state information of the system;

[0103] The artificially intelligent model used in the cognitive decision-making step is a deep reinforcement learning model. The steps include:

[0104] Sp3.1. Take the network cooperation information received from other network nodes and the security state information of the system as additional state inputs to the deep reinforcement learning model;

[0105] Sp3.2. Train or guide the deep reinforcement learning model by jointly optimizing the reward function for multiple communication objectives;

[0106] Sp4. Use the Adaptive Transceiver Control Module (ATCM) to configure the communication parameters of the physical layer (PHY) and the media access control layer (MAC) of the system according to the control instructions to perform adaptive adjustment across multiple communication parameter domains;

[0107] Distributed Cooperation Step: Use the distributed cooperation module to share interference information and link quality information with other communication nodes in the network, and use the obtained information for the cognitive decision-making step;

[0108] Endogenous Security Enhancement Step: Use the endogenous security enhancement module to execute at least one security mechanism among dynamic heterogeneous redundancy, mimic defense, or physical layer authentication according to the instructions generated in the cognitive decision-making step or preset policies. Specific Embodiment Three:

[0110] Based on the technical solutions of Specific Embodiment 1 and Specific Embodiment 2, a practical application case is further given to illustrate:

[0111] In countries where natural disasters such as earthquakes and typhoons occur frequently, establishing a stable and reliable emergency communication network in the first time after the disaster is crucial for saving lives and coordinating rescue efforts. However, the disaster area is often accompanied by damaged infrastructure, power outages, spectrum chaos, and poor channel conditions, posing severe challenges to traditional communication systems. The deployment of the "Intelligent Sensing and Interference Resistance" Cognitive Cooperative Adaptive Resilient Carrier Communication System (CCARCS) is precisely to provide unprecedented communication guarantee capabilities in such an extremely complex electromagnetic environment. Its breakthrough lies in the unprecedented deep integration of deep environment perception, AI cognitive decision-making, multi-domain adaptive execution, distributed intelligent cooperation, and endogenous proactive security.

[0112] In practical applications, the CCARCS systems deployed on nodes such as command vehicles, individual soldier terminals, and drones first use their Environment Sensing Module (ESM) to precisely "portray" the chaotic spectrum in the disaster area. This far exceeds the simple energy detection of traditional spectrum analyzers. The ESM, using its broadband monitoring capabilities and multi-dimensional feature extraction technologies (spectrum, time, space, cyclostationarity, etc.), can finely distinguish broadband impulse noise from damaged power facilities, spurious radiation generated by temporary generators, intermodulation interference among a large number of rescue devices of different systems (even including old analog radios), and atypical signals that may be caused by atmospheric disturbances or special geological activities. Its breakthrough lies in the built-in AI analysis unit: pre-trained deep learning models such as CNN + LSTM can quickly identify these known or similar-pattern interference types, while anomaly detection algorithms such as autoencoders can "pick out" previously unseen unknown interference signals from the background and mark them, which is difficult to achieve by traditional rule- or template-matching methods. Combining the DOA estimation of multi-antenna arrays, the ESM generates a high-precision, dynamically updated "interference map" (including frequency, power, direction, type, confidence, etc.), providing an unprecedented, comprehensive and accurate situation information basis for the intelligent decision-making of the entire system.

[0113] Based on this precise "battlefield intelligence", the Cognitive Decision-making Engine (CDE), as the intelligent core of the system, demonstrates its revolutionary autonomous decision-making ability. It receives the situational information from ESM, combines the local terminal status (such as remaining battery power, priority of data to be sent), key interference intelligence and link quality reports shared from other rescue nodes via the Distributed Cooperative Module (DCM), and the security situation reported by the Endogenous Security Enhancement Module (ESEM), and uses its Deep Reinforcement Learning (DRL) models (such as PPO / SAC) for global thinking. This AI-driven decision-making process is a real breakthrough. It no longer relies on rigid preset rules, but based on the learned experience, in the complex multi-dimensional state space, it can generate an optimal combination of control instructions covering the entire communication parameter domain (frequency, time, space, power, waveform, coding, modulation) in real-time and autonomously. For example, when a high-priority vital sign data transmission request arrives, its built-in multi-objective reward function will automatically increase the weights of reliability (such as selecting a strong error-correcting code) and low latency, even at the cost of sacrificing a certain throughput or increasing instantaneous power consumption; while when transmitting ordinary disaster situation images, it may focus more on spectral efficiency and energy conservation.

[0114] The intelligent instructions of CDE are then quickly and precisely executed by the Adaptive Transceiver Control Module (ATCM), which is the key link to transform decisions into physical layer resilience. The innovative value of ATCM lies in its extreme flexibility and multi-dimensional linkage ability. It can not only adjust the rate, but also perform in-depth physical layer reconstruction. For example, when CDE determines that narrowband interference is severe and adjacent spectra are available, ATCM will switch to the FBMC waveform with stronger spectral shaping ability to avoid interference; when it determines that the channel has severe fast fading due to building blockage and movement, it will select a coding and modulation scheme with better diversity performance and may instruct the multi-antenna system to perform Space-Time Block Coding (STBC) or receive diversity combining; when ESM accurately locates the direction of a strong interference source, ATCM will calculate and apply adaptive beamforming weights to form a deep spatial null in the receiving end in that direction, or precisely align the energy to the target receiving node at the transmitting end, greatly improving the signal-to-noise ratio of a specific link. At the same time, according to the real-time channel quality, dynamically adjust the cyclic prefix (CP) length of OFDM to adapt to the changing multipath environment, select LDPC, Polar or even Fountain codes and finely adjust the code rate to achieve the best balance between rate and reliability.

[0115] The resilience at the network level is provided by the Distributed Cooperative Module (DCM). Its core innovation lies in achieving swarm intelligence and resource sharing. Rescue teams are widely distributed, and the perception and decision-making capabilities of individual nodes are limited. Through the DCM, nodes can efficiently share key interference information summaries and link states. Furthermore, based on the Federated Learning (FL) mechanism, all nodes can collaboratively train an interference recognition model or a channel prediction model adapted to the specific electromagnetic environment of the current disaster area without exposing their local raw data, significantly improving the accuracy and timeliness of global situation awareness, which is incomparable to traditional distributed systems. At the same time, the DCM supports a distributed resource negotiation protocol, enabling nodes of different rescue teams to intelligently coordinate the use of spectrum, time slots, and power, avoiding "self-consumption" mutual interference and maximizing the overall communication efficiency of the network with limited resources. This collaborative ability ensures that even in the case of partial node failures or isolations, the network can still maintain the highest degree of connectivity and service capabilities.

[0116] Finally, the Endogenous Security Enhancement Module (ESEM) provides a solid security foundation for rescue communication. Its breakthrough lies in deeply integrating security capabilities into the communication system itself to achieve active defense. In the chaotic disaster area environment, preventing malicious interference, information theft, or identity forgery is crucial. The physical layer authentication function of the ESEM uses the unique channel or radio frequency "fingerprint" of each device for identity verification, effectively preventing the attempt of forged devices to access the network. In the face of potential malicious interference or signal analysis attacks, the CDE can instruct the ESEM to initiate mimic defense. Through the high-speed, pseudo-random multi-dimensional parameter (frequency, time, coding, waveform micro-parameters, etc.) hopping driven by the CSPRNG, it makes the communication signal difficult to be tracked, locked, or decoded, greatly improving the survivability. At the same time, the Dynamic Heterogeneous Redundancy (DHR) mechanism provides backups for critical processing links. Even if a certain software implementation is attacked, the system can switch to other heterogeneous implementations to ensure that the core functions are not interrupted.

[0117] In summary, through the collaborative application of a series of breakthrough technologies such as deeply integrated intelligent perception, AI cognitive decision-making, multi-domain adaptive execution, distributed cooperation, and endogenous security, the CCARCS system can build an emergency communication network with far superior capabilities to traditional communication systems, truly possessing high resilience, high efficiency, and high security in the extremely complex electromagnetic environment after large-scale disasters such as earthquakes, providing strong technical support for saving lives and coordinating rescue efforts. Specific Embodiment Four:

[0119] Based on the technical solutions of Specific Embodiment One and Specific Embodiment Two, a practical application case is further given for illustration:

[0120] Autonomous driving technologies, especially highly autonomous driving at L4 / L5 levels and platoon operation, pose unprecedented requirements for ultra-high reliability, ultra-low latency (URLLC), and high security in vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications. In high-density urban environments with numerous high-rise buildings, congested wireless signals, and potential malicious interference, traditional V2X communication technologies face significant challenges. The application of the "Intelligent Sensing and Interference Resistance" Cognitive Cooperative Adaptive Resilient Carrier Communication System (CCARCS) aims to provide a revolutionary communication solution for autonomous driving platoons that can ensure safe and efficient cooperation in such a harsh environment. Its innovative value lies in its intelligent adaptation ability to handle extremely complex environments and the embedded proactive safety protection mechanism.

[0121] The CCARCS system deployed on each vehicle in the platoon, with its Environment Sensing Module (ESM), acts as an "electronic sentry". It not only monitors the V2X dedicated frequency band (such as 5.9 GHz) but also scans the surrounding ISM frequency bands that may cause interference. Its innovation lies in not only being able to "see" interference but also "understand" it. In high-density urban areas, the ESM can utilize its fine-grained spectrum analysis and multi-dimensional feature extraction capabilities to distinguish complex interference signals from densely deployed Wi-Fi / Bluetooth devices, other vehicle sensors (radar, Lidar), mobile base stations, and even power lines or traffic signal controllers. Through AI classification models (such as CNN + LSTM), it can identify the patterns of these common interference sources. More importantly, in the face of the severe multipath and NLOS environment caused by the urban canyon effect, the ESM can accurately estimate the multipath delay spread and fast time-varying characteristics (Doppler frequency shift) of the channel. Combining DOA estimation, it can also preliminarily locate the main interference sources (such as a fixed strong interference transmitter or an abnormal vehicle in the vicinity). The anomaly detection algorithm (Autoencoder) is responsible for identifying those atypical and possibly malicious signals, such as GPS spoofing signals or interference attack signals targeting V2X. All this information is integrated into a dynamic interference map in real time to provide a basis for decision-making.

[0122] The core of the system - the Cognitive Decision Engine (CDE) - demonstrates its AI decision-making intelligence tailored for autonomous driving scenarios. It receives the precise environmental portrait from the ESM and combines it with the vehicle's sensor data (speed, acceleration, steering angle), driving intention, high-precision positioning information, and key cooperative information obtained from other vehicles in the fleet through the DCM (such as neighboring vehicle status, planned trajectory, link quality), and also considers the safety status of the ESEM. Its breakthrough lies in that the DRL model (PPO / SAC) of the CDE is trained to prioritize URLLC requirements. Its multi-objective reward function gives extremely high rewards for control instruction transmissions that meet millisecond-level latency and "five nines" or even higher reliability, and imposes severe penalties on any communication failures that may endanger safety. Based on this, the CDE can make highly context-aware, predictive, and cross-domain optimal communication decisions. For example, when predicting that it is about to enter an intersection with known severe interference, the CDE will instruct the ATCM to switch to the most robust communication mode in advance; during high-speed platooning, adjust the waveform parameters according to real-time Doppler estimation; when detecting a potential interference attack, cooperate with the ESEM to enhance the safety level.

[0123] The Adaptive Transceiver Control Module (ATCM) is responsible for translating the intelligent decisions of the CDE into extremely flexible and fast-responsive physical layer actions, which is the key execution link for ensuring URLLC. Its innovative value lies in its deep, context-driven adaptive capabilities: when the CDE determines that the channel is in a severe NLOS and fast fading state, the ATCM may not only reduce the rate but also switch to a coding scheme with stronger diversity capabilities (such as low code rate Polar codes) and enable Space-Time Block Coding (STBC); when it is necessary to resist strong co-channel interference, the ATCM uses a multi-antenna array to perform precise adaptive beamforming and spatial nulling to suppress interference outside the receiving end; to counteract urban multipath, dynamically adjust the CP length of OFDM; to maximize the transmission success rate of short-time burst data, finely adjust the MCS level and HARQ retransmission strategy. This real-time, joint, and intelligent optimization in multiple dimensions such as waveform, coding, modulation, and spatial processing is unparalleled by traditional fixed-parameter or limited adaptive systems.

[0124] The Distributed Cooperative Module (DCM) plays a crucial role in the fleet scenario. Its breakthrough lies in achieving efficient, reliable, and intelligent fleet-level cooperation. Vehicles need to frequently exchange a large amount of status and control information with extremely low latency. DCM not only ensures the reliable transmission of this information but also enables more advanced cooperation: by sharing their respective interference perception results (possibly through federated learning to jointly train an interference prediction model adapted to the current urban area), the fleet can form a unified understanding of the electromagnetic environment and make better collective decisions (such as uniformly switching to a certain clean frequency band). By executing a distributed resource negotiation protocol optimized for low latency, vehicles within the fleet can avoid mutual interference in V2V communication and ensure unobstructed priority transmission channels for critical control information. This swarm intelligence significantly improves the operating efficiency and safety of the entire fleet in complex environments.

[0125] Finally, the Endogenous Security Enhancement Module (ESEM) provides indispensable proactive security protection for autonomous driving fleet communication. Its innovation lies in internalizing defense as an inherent property of the system. In high-risk autonomous driving applications, preventing malicious manipulation is crucial. The physical layer authentication mechanism of ESEM ensures the authenticity of the message source by verifying the "physical fingerprints" of vehicle-to-vehicle communication, effectively resisting attacks such as forged identities and message injection. When the system senses targeted interference or GPS spoofing attacks, the CDE will instruct ESEM to initiate mimic defense, making it difficult for attackers to lock and disrupt the critical fleet cooperation communication link by performing unpredictable high-speed parameter jumps in dimensions such as frequency, time, and coding. Dynamic Heterogeneous Redundancy (DHR) provides additional fault tolerance for software modules that process critical security information and control instructions.

[0126] The CCARCS system provides a breakthrough communication solution for the cooperative operation of autonomous driving fleets in high-density urban environments through the deep integration of its unique intelligent perception, AI cognitive decision-making, multi-domain adaptive execution, distributed cooperation, and endogenous security technologies. It not only solves the reliability and latency pain points of traditional V2X technologies in complex electromagnetic environments but also significantly enhances the system's anti-attack ability through the embedded proactive security mechanism, laying a solid communication foundation for realizing safe and efficient future intelligent transportation.

[0127] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] The above has introduced in detail the technical solutions provided by the embodiments of the present invention. Specific examples are used herein to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention. At the same time, for those of ordinary skill in the art, based on the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0129] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a reference structure" does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0130] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An adaptive interference suppression carrier communication system for complex electromagnetic environments, characterized in that: The described adaptive interference suppression carrier communication system includes the following components: An environment perception module, which is used to: monitor electromagnetic signals within a preset frequency band and extract multi-dimensional features of the electromagnetic signals; the multi-dimensional features include one or more of spectral features, time-domain features, and spatial-domain features; A cognitive decision-making engine, which is used to receive environmental situation information and the internal state information of the adaptive interference suppression carrier communication system; Multiple communication parameter domains, which include two or more of the frequency domain, time domain, spatial domain, power domain, waveform domain, coding domain, and modulation domain; An adaptive transceiver control module, which is used to receive control instructions; according to the control instructions, configure the communication parameters of the physical layer or the media access control layer of the system; A distributed cooperation module, which is used to perform information interaction with other communication nodes in the network; An in-built security enhancement module, which is used to execute in-built security mechanisms according to the instructions of the cognitive decision-making engine or preset security policies.

2. The adaptive interference suppression carrier communication system based on a complex electromagnetic environment according to claim 1, wherein The spatial-domain features extracted by the environment perception module contain signal arrival direction information, and the environmental situation information is constructed into an interference map data structure, which contains information on the frequency, power, direction, and type of interference signals.

3. The adaptive interference suppression carrier communication system based on a complex electromagnetic environment according to claim 1, wherein The cognitive decision-making engine uses a pre-trained artificial intelligence model as a deep reinforcement learning model; and the state space input of the deep reinforcement learning model further includes network cooperation information received from other network nodes and the security state information of the system; the reward function of the deep reinforcement learning model jointly optimizes communication throughput, bit error rate, latency, power consumption, and security metrics.

4. The adaptive interference suppression carrier communication system based on a complex electromagnetic environment according to claim 1, wherein The distributed cooperation module is further configured to: participate in distributed model training based on federated learning, where multiple nodes cooperate to train at least one shared artificial intelligence model for interference recognition, channel prediction, or policy generation, and the original perception data of the nodes does not leave the local nodes during the training process, execute a distributed resource negotiation protocol, and cooperate with other nodes to perform communication resource allocation or interference suppression operations.

5. The adaptive interference suppression carrier communication system based on a complex electromagnetic environment according to claim 1, characterized in that The communication parameter adjustment configured by the adaptive transceiver control module (ATCM) includes at least one of the following: Select between multiple candidate waveforms according to the environmental situation information or service requirements, and adaptively adjust the key parameters of the selected waveform; Select between multiple channel coding schemes according to the environmental situation information, and adaptively adjust the coding rate and modulation method; Control the multi-antenna system to perform adaptive beamforming or spatial nulling according to the environmental situation information.

6. The method of the adaptive interference suppression carrier communication system based on the complex electromagnetic environment according to any one of claims 1-5, characterized in that, The adaptive interference suppression carrier communication method includes the following steps: Sp1. Use the environment perception module to monitor electromagnetic signals within a preset frequency band, extract multi-dimensional features of the electromagnetic signals, and generate environmental situation information containing interference information and channel state information. The environmental situation information generated by the environment perception module contains signal arrival direction information and is constructed into an interference map data structure; Sp2. Identify the type of interference signal or indicate unknown interference using a pre-trained classification model or anomaly detection algorithm, and include the result in the environmental situation information; Sp3. Cognitive decision-making step: Use the pre-trained artificial intelligence model in the cognitive decision-making engine to generate control instructions covering multiple communication parameter domains based on the environmental situation information and the internal state information of the system; Sp4. Use the adaptive transceiver control module to configure the communication parameters of the physical layer and the media access control layer of the system according to the control instructions to perform adaptive adjustment across the multiple communication parameter domains.

7. The adaptive interference suppression carrier communication method based on the complex electromagnetic environment according to claim 6, characterized in that The artificial intelligence model used in the cognitive decision-making step is a deep reinforcement learning model, and this step further includes: Sp3.

1. Take the network cooperation information received from other network nodes and the security state information of the system as additional state inputs to the deep reinforcement learning model; Sp3.

2. Train or guide the deep reinforcement learning model by jointly optimizing the reward function for multiple communication objectives.

8. The adaptive interference suppression carrier communication method based on a complex electromagnetic environment according to claim 6, wherein The method further includes: Distributed cooperation step: Use the distributed cooperation module to share interference information and link quality information with other communication nodes in the network, and use the obtained information for the cognitive decision-making step; Endogenous security enhancement step: Use the endogenous security enhancement module to execute at least one security mechanism of dynamic heterogeneous redundancy, mimic defense, or physical layer authentication according to the instructions generated in the cognitive decision-making step or preset policies.

Images