【作者】 周兰；

【导师】 陈文； 杨韬；

【作者基本信息】 四川大学 ， 电子信息（专业学位）， 2023， 硕士

【摘要】 无线传感器网络(Wireless Sensor Networks,WSNs)作为迅速发展的无线通信技术,广泛应用于战场环境感知、自然环境监测、智能电网、工业控制等领域。WSNs可通过传感器节点收集和处理数据并传输至用户。在WSNs中,传感器节点拥有有限的计算和存储资源、能源和带宽,且通常运行在不受保护且无人监管的环境中。分布式检测是WSNs中的一种重要的数据融合方法,核心思想是在目标区域部署多个传感器以测量目标的各种物理状态。传感器基于自身观测结果进行独立的局部决策,然后通过无线信道将决策发送给合法融合中心(Ally Fusion Center,AFC)。AFC融合所有收到的数据,做出全局的最终决策。然而,由于传感器数据采用广播式传输,恶意攻击者可轻易窃听传感网络的数据,威胁数据传输机密性。此外,WSNs中传感器通常依赖电池供电,限制了计算、存储和续航能力。因此,传统需要复杂数学计算的加密方法难以适用于资源受限的WSNs环境。针对单跳WSNs分布式数据融合中,传统的加密方案因软硬件资源限制而难以适用的问题,本文提出了基于信息隐藏和物理层加密的轻量级安全传输方案。相较于传统加解密算法,本文方法仅依赖数据的01随机翻转实现数据的物理层加密,在保障传输数据安全性的同时,降低了计算代价,更加适用于资源受限的WSNs环境。本文的方法具有较好的抗干扰性,具有抵抗信道环境变化和敌方融合中心(Enemy Fusion Center,EFC)恶意攻击的能力,此外,采用信息隐藏方式保护传感器数据,增加了安全性。在远距离多跳自组网场景中部署众多传感器时,大部分传感器需采用中继多跳传输方式,利用节点转发与AFC通信。多跳网络中由于远距离物理信道无法长时间保持状态稳定,难以利用传统的物理层加密技术实现AFC与传感器间的加密传输。本文提出了一种基于拓扑结构和压缩感知加密技术的分布式数据安全融合方案。该方案将网络拓扑结构特征引入到压缩感知矩阵的构造中,并采用压缩感知加密技术实现节点间的安全数据传输。通过仿真实验和真实传感网络下的物理实验,本文验证了该方案在多跳WSNs环境下能有效地抵御窃听攻击和EFC攻击,保障网络的数据安全。此外,通过评估方案的能量消耗,证实了本文的方法较传统加密方法功耗显著降低,更适用于资源受限的WSNs环境。综上所述,本文提出的基于信息隐藏和物理层加密的轻量级安全传输方案以及基于拓扑结构和压缩感知加密技术的安全传输方案,为WSNs的分布式状态估计信息融合提供了可靠保障。该方案既降低了能量消耗和计算负担,又能够保证WSNs数据传输的安全性。通过对本文方案的研究和实验验证,为无线传感器网络数据传输安全探索了新的技术途径。更多还原

【Abstract】 Wireless Sensor Networks(WSNs)are rapidly evolving wireless communication technologies used in battlefield environmental awareness,natural environment monitoring,smart grids,industrial control,etc.WSNs collect and process data from sensor nodes and transmit it to users.In WSNs,sensor nodes have limited computing and storage resources,energy and bandwidth,and often operate in unprotected and unsupervised environments.Distributed detection is an important approach to data fusion in WSNs,with the core idea being to deploy multiple sensors in a target area to measure various physical states of the target.The sensors make independent local decisions based on their own observations,which are then sent over a wireless channel to an Ally Fusion Center(AFC),which fuses all the data received to make a global final decision.However,because sensor data is broadcast in the air,a malicious attacker can easily eavesdrop on the sensing network,threatening the confidentiality of the data transmission.In addition,sensors in WSNs typically rely on battery power,limiting computation,storage and endurance.As a result,traditional encryption methods that require complex mathematical calculations are difficult to apply in resource-constrained WSNs environments.To address the problem that traditional encryption schemes are difficult to apply in single-hop WSNs’ distributed data fusion due to hardware and software resource constraints,a lightweight secure transmission scheme based on information hiding and physical layer encryption is proposed.Compared with traditional encryption and decryption algorithms,the proposed scheme only relies on the randomly flipping of binary data to achieve physical layer encryption of data,which is more suitable for resource-constrained WSNs environments as it reduces the computational cost while safeguarding the security of transmitted data.The proposed scheme has good resistance to interference,channel environment changes and Enemy Fusion Center(EFC)malicious attacks,in addition,the use of information hiding method to protect the sensor data and increase data security.When deploying numerous sensors in a long-range multi-hop Ad Hoc scenario,most sensors need relayed multi-hop transmissions to communicate with the AFC using node forwarding.In multi-hop networks,it is difficult to achieve encrypted transmission between AFCs and sensors using traditional physical layer encryption techniques because the long-range physical channel cannot maintain state stability for a long time.A distributed data security fusion scheme based on topology and compressed sensing encryption technology is proposed.The scheme incorporate network topology features into the construction of a compressed sensing matrix and uses compressed sensing encryption to achieve secure data transmission between nodes.Through simulation experiments and physical experiments under real sensing networks,it is verified that the scheme can effectively resist eavesdropping attacks and EFC attacks in a multi-hop WSNs environment and safeguard the data of the network.In addition,by evaluating the energy consumption of the scheme,it is confirmed that the proposed scheme consumes significantly less power than traditional encryption methods and is more suitable for resource-constrained WSNs environments.In summary,the lightweight secure transmission scheme based on information hiding and physical layer encryption,as well as the secure transmission scheme based on topology and compressed sensing encryption techniques proposed in this thesis,provide a reliable guarantee for distributed state estimation information fusion in WSNs.The scheme reduces the energy consumption and computational burden,while it is also able to guarantee the security of WSNs data transmission.The research and experimental validation of the scheme in this thesis explored new technical approaches for data transmission security in WSNs.更多还原