针对强噪声环境下雷达新型有源干扰识别准确率不高的问题,提出了一种KPCA-SAE-BP网络算法。提取干扰信号时域、频域、波形域、小波域、双谱域等特征构建67维输入空间,经过核主成分分析(kernel principal component analysis,KPCA)将高...针对强噪声环境下雷达新型有源干扰识别准确率不高的问题,提出了一种KPCA-SAE-BP网络算法。提取干扰信号时域、频域、波形域、小波域、双谱域等特征构建67维输入空间,经过核主成分分析(kernel principal component analysis,KPCA)将高维数据进行非线性降维与重构,利用SAE-BP神经网络完成分类识别。仿真结果表明,在干噪比(JNR)大于-1 dB的强噪声环境中,KPCA-SAE-BP网络算法对6种新型有源干扰的识别准确率达到90%以上,训练与识别时间少于0.7 s。相同参数条件下,与经典BP神经网络、SAE-BP网络、KPCA-BP网络、GA-BP网络相比,具有更好的检测识别性能。展开更多
Based on a review of 28 Horizon Europe-funded CCAM projects, this paper studies the current state of Connected, Cooperative, and Automated Mobility (CCAM) and identifies significant research gaps in taxonomy, cybersec...Based on a review of 28 Horizon Europe-funded CCAM projects, this paper studies the current state of Connected, Cooperative, and Automated Mobility (CCAM) and identifies significant research gaps in taxonomy, cybersecurity, Artificial Intelligence (AI) and 6G research, that hinder the advancement of a future-ready CCAM infrastructure. The research emphasizes the crucial role of infrastructure in achieving autonomous mobility, shifting focus from the current vehicle-centric approach. It critiques the SAE J3016 taxonomy for its lack of emphasis on infrastructure and proposes an updated framework with an automation level dedicated to infrastructure automation. The paper highlights the existential threats posed by Quantum Computers (QC) and AI, stressing the need for quantum-safe cybersecurity measures and an ethical, controllable AI framework proposing a decentralized Collective Artificial Super Intelligence (CASI) framework. Identifying the critical need for a cooperative approach involving Road and Transport Authorities (RTAs) to achieve 100% vehicle connectivity and robust digital infrastructure, the study outlines the European Commission’s Vision 2050 goals, aiming for zero fatalities, zero emissions, and sustainable mobility. The paper concludes by providing recommendations for future research directions to accelerate the development of a comprehensive, secure, and efficient CCAM ecosystem.展开更多
文摘Based on a review of 28 Horizon Europe-funded CCAM projects, this paper studies the current state of Connected, Cooperative, and Automated Mobility (CCAM) and identifies significant research gaps in taxonomy, cybersecurity, Artificial Intelligence (AI) and 6G research, that hinder the advancement of a future-ready CCAM infrastructure. The research emphasizes the crucial role of infrastructure in achieving autonomous mobility, shifting focus from the current vehicle-centric approach. It critiques the SAE J3016 taxonomy for its lack of emphasis on infrastructure and proposes an updated framework with an automation level dedicated to infrastructure automation. The paper highlights the existential threats posed by Quantum Computers (QC) and AI, stressing the need for quantum-safe cybersecurity measures and an ethical, controllable AI framework proposing a decentralized Collective Artificial Super Intelligence (CASI) framework. Identifying the critical need for a cooperative approach involving Road and Transport Authorities (RTAs) to achieve 100% vehicle connectivity and robust digital infrastructure, the study outlines the European Commission’s Vision 2050 goals, aiming for zero fatalities, zero emissions, and sustainable mobility. The paper concludes by providing recommendations for future research directions to accelerate the development of a comprehensive, secure, and efficient CCAM ecosystem.
