USB移动存储设备中异常病毒数据检测技术研究

Abnormal Virus Data Detection Technology Research in USB Mobile Storage Device

对USB移动存储设备中的异常病毒数据进行检测,可以延长USB移动存储设备寿命,提高数据利用率,减少系统运行时间;当前方法利用轨迹点片段异常对USB移动存储设备中的异常病毒数据进行检测,将几个独立的USB移动存储设备中的异常病毒数据属性进行结合,针对现有的异常病毒数据点的异常轨迹进行检测,以病毒数据的位置、速度以及方向为检测对象;该方法对USB移动存储设备中的异常病毒数据检测效率低,不适用于大规模的USB移动存储设备中的异常病毒数据检测;为此,提出一种基于PATRICIA树的USB移动存储设备中异常病毒数据检测方法;该方法利用K-means算法将USB移动存储设备中的数据划分为K个类,并利用欧几里德距离对各个类间的相似度进行衡量,然后在独立分量分析的基础上加入遗忘因子,对USB移动存储设备中异常病毒数据检测时含噪进行测量估计,最后利用小波分析法,通过设置USB移动存储设备中异常病毒数据判定阈值,以及标准化以后的小波系数内绝对值,与判定阈值的比较完成病毒数据的检测;实验结果证明,所提的检测方法可以高精度地对USB移动存储设备中异常病毒数据进行检测,更加符合该领域发展实际意义。

The detection of abnormal virus data in the USB mobile storage device to can extend the life of USB mobile storage device, im- prove the data utilization and reduce the running time of the system. Segments of current method using trajectory point anomalies of USB re- movable storage device detect virus data, several independent USB removable storage device when the virus data attributes, in view of the ex- isting exception virus detect abnormal data points of trajectory, virus data position, speed and direction as test object. This method is less ef- ficient in detecting abnormal virus data in USB mobile storage devices, and does not apply to the detection of abnormal virus data in a large number of USB mobile storage devices. To this end, an abnormal virus data detection method is proposed in a USB mobile storage device based on the PATRICIA tree. The method using the K - means algorithm to data in USB removable storage device is divided into K classes, and by using Euclidean distance to measure the degree of similarity between the classes, and then on the basis of the independent component analysis to join the forgetting factor, abnormal data virus detection of USB removable storage device when the measured signals with noise es- timation, the use of wavelet analysis, by setting the USB removable storage device in abnormal virus data to determine the threshold, and standardized the wavelet coefficient absolute value, compared with the decision threshold to complete the data of the virus. The experimental results show that the proposed detection method in this paper can accurately for USB mobile storage devices, abnormal virus data for testing, more in line with the development of practical significance in this field.