Research On Smart Information Security Detection Technologies Towards Defense-in-Depth System Of Internet Of Battlefield Things

The key to achieving"detection before the enemy,strike before the enemy"in mod-ern warfare lies in information superiority-In other words,we can obtain and maintain a more comprehensive and accurate situational awareness than the enemy in combat.Internet of Battlefield Things(Io BT)is an Internet of Things(Io T)system used in the battlefield environment.Its primary function is to assist in tactical synergy tasks such as jamming,trajectory,and fire synergy through transmitting perceived battlefield situa-tional sense information,combat instructions,and other two-way information to form a war advantage.With the development of Artificial Intelligence(AI)and command and control technology,more and more Io BT equipment represented by unmanned aerial vehicles,unmanned combat vehicles,and communication terminals has begun to ac-cess the Io BT.Io BT becomes the core base unit for building a complete C~4ISR system.Compared with traditional battlefield information systems,the number of equipment accessed by the Io BT has increased exponentially,and so does the reliance of the com-bat system on the network.This new trend has largely broadened the attack surface of the Io BT.The traditional information security strategy that relies on strict network boundary control can no longer meet the security needs of the Io BT.Constructing an effective defense-in-depth system for Io BT information security is of great significance to ensure the efficiency and reliability of command and intelligence.However,most Io BT research at this stage mainly focuses on the technical imple-mentation level,and research on information security is relatively scarce.Besides,mul-tiple reasons make it difficult to embed security capabilities into the Io BT devices,such as tangled sources of Io BT devices,complicated technical functions,strict confidential-ity requirements,and the absence of security personnel.Starting from the actual sce-nario of Io BT,this thesis promotes the construction of the information security defense system of Io BT by researching novel smart information security detection technologies.It focuses on the security needs that have not been met at the device,communication,service,and data levels.At the device level,this thesis designs a new software supply chain vulnerability detection and software component analysis solution based on Pseudo-code and Graph Neural Network(GNN)for the firmware of Io BT equipments,PG-SCEye.It performs software component analysis based on cross-architecture function-level vulnerability detection,effectively assisting security engineers in long-term Io BT security manage-ment.Experiments show that the vulnerability detection accuracy of PG-SCEye ex-ceeds 99%,which is a significant improvement over existing works.At the communication level,this thesis designs Io BTGuard Eye,a novel malicious traffic detection approach based on Bi-directional Long Short-Term Memory(Bi LSTM)network for the Application Programming Interface(API)commonly used in Io BT de-vices.Io BTGuard Eye is the first to introduce a multi-model flow-splitting design into malicious traffic detection.The detection rate of malicious traffic for a specific API exceeds 99%,which has significant advantages over existing approaches,indicating its superiority for critical Io BT APIs with distinctive traffic features.At the service level,this thesis designs a novel real-time detecting and blocking approach named SQLState Guard for SQL Injections which is the most data-hazardous attack type.Unlike the traditional Web Application Firewall(WAF),this thesis in-troduces Runtime Application Self-Protection(RASP)idea into the middleware.It achieves statement-level non-intrusive real-time SQL Injection detecting and blocking based on an improved Bi LSTM network.Experiments show that the detection accuracy of SQLState Guard for SQL Injections exceeds 99%.It can identify the technical and tactical methods of SQL Injections with high accuracy.Compared with existing meth-ods based on WAF and traffic-level attack detection,this method has an overwhelming advantage.At the data level,this thesis conducts a series of offensive and defensive counter-measures against the authenticity detection of voice signaling in the Io BT.In terms of attacks,Si F-Deep VC is introduced for voice cloning attacks that can effectively bypass existing detection systems without affecting expression based on speaker-irrelative fea-tures.Experiments show that existing detection systems believe the voices generated by Si F-Deep VC are"more human than human".After verifying the effectiveness of the attack framework,this thesis designs a high-robust AI-synthesized speech detection network named AD-Net for combating voice cloning attacks,and its effectiveness has been demonstrated through experiments.In summary,this thesis’s smart information security detection research endoge-nizes security capabilities into the Io BT system level-by-level,alleviating the contra-diction between complex security management,lagging security response,offensive and defensive asymmetries,and the safe operation of Io BT systems.Thus,this the-sis aims at providing in-depth information security detection capabilities for Io BT in device security,communication security,service security,and data security without in-tervening in equipment development.