摘要
改进了一种基于神经网络吸引子计算的无线传感器公钥加密算法,通过引入改进的q-composite密钥分发机制替代原有的Differ-Hellman密钥协商算法,利用其对神经网络初始化从而产生加密数据的吸引子。在对明文加密之前使用吸引子Markov模型编码,最后将吸引域作为密文输出从而达到加密效果。在OMNet++与Matlab环境下对算法进行了仿真并与同类算法在能耗方面进行了对比。对算法的雪崩效应,密文平衡性及独立性进行了测试,并给出了算法的安全性、扩展性和抗破译能力相关分析。
Improve a new kind of public key encryption algorithm for wireless sensor network based on attractor computation in neural network. Through improved q-composite key distribute scheme to replace original Differ-HeUman key agreement algorithm, synaptic matrix of neural network is build. Adopting a kind of Markov Process to code the attractors then select the attraction domain as output cipher text. Execute avalanche, cipher-text balance and independence test to the algorithm in OMNet＋＋ and Matlab. Finally we analyzed the algorithm＇s security and scalability theoretically.
