论文部分内容阅读
为了保证智能手机的信息安全,提出一种多传感器信息融合的手机用户身份认证方法。首先从手机用户操作手势的姿态角变化、运动幅度以及旋转程度等方面提取信号特征,然后分别以三类单特征的动态时间规整识别结果作为独立证据构造基本概率分配函数,进而采用Dempster/Shafer证据理论对其进行融合。为缓解证据间出现的冲突,避免其影响融合效果,引入加权证据方法。通过计算各证据之间的相似性来衡量证据间的冲突程度进而确定各证据的可信度,并且对各证据进行加权修正以降低可信度小的证据对融合结果的影响,最后根据融合结果做出决策。仿真结果表明,该算法性能优于对比算法,能有效对手机用户进行身份识别。
In order to ensure the information security of smart phones, this paper proposes a mobile phone user authentication method based on multi-sensor information fusion. Firstly, the signal features are extracted from the change of gesture angle, movement amplitude and rotation degree of the hand gesture of the mobile phone users. Then, the basic probabilistic distribution function is constructed based on the dynamic time warping results of three types of single features respectively, and then Dempster / Shafer evidence Theory of its integration. In order to alleviate the conflict between the evidences and avoid affecting the fusion effect, we introduce the weighted evidence method. By calculating the similarity between the evidence to measure the degree of conflict between the evidence and then determine the credibility of the evidence, and the weight of the evidence to be corrected to reduce the credibility of the evidence of the fusion results, and finally based on the results of the fusion Make a decision. Simulation results show that the proposed algorithm outperforms the contrast algorithm and can effectively identify mobile phone users.