摘要
为提高无线传感器网络的感知精度,提出了一种基于提升小波变换和自适应多项式拟合的多模数据压缩算法AMLP(Adaptive Multiple-Modality Data Compression Algorithm Based on Lifting Wavelet and Adaptive Polynomial Fitting)。在给定相关度阈值的前提下,AMLP算法先对数据进行灰色关联聚类,再对类中的相关数据进行自适应的多项式拟合,然后把未拟合的特征数据抽象成一个矩阵,利用提升小波变换去除数据的时间和空间相关性。最后,通过游程编码对数据作进一步压缩。仿真结果表明,AMLP算法能够有效去除不同数据间的冗余信息以及同种数据间的时间和空间冗余信息,提高压缩比,降低网络能耗。与基于小波的自适应多模数据压缩算法AMMC(Adaptive Multiple-Modality Data Compression Algorithm Based on Wavelet)相比,AMLP算法的数据恢复精度大大优于AMMC算法,压缩比和能耗相近。因此,AMLP算法更适用于要求高精度数据的传感器网络应用,如地质灾害监测、医疗和军事领域。
In order to improve the accuracy of the sensed data in wireless sensor networks, we propose an adaptive multiple-modality data compression algorithm based on lifting wavelet and adaptive polynomial fitting（AMLP） in this paper. The AMLP algorithm clustes all sensed data by using grey clustering technique, and approximates the relevant data in each cluster by using adaptive polynomial fitting. After that, the un-fitted characteristic data is abstracted as a matrix and then the AMLP algorithm uses lifting wavelet to remove the temporal and spatial redundancy. Finally, it compresses the data by RLE （Run Length Encoding）coding method. Simulation results demonstrate that the AMLP algorithm can effectively remove the redundancy both within the values of a single measurement as well as among values of different measurements, thus to decrease the energy consumption of sensor node. Although the proposed al- gorithm does not have advantage in compression ratio and energy consumption when compared to the adaptive multiple-modality data compression algorithm based on wavelet（AMMC）, but it significantly outperforms the AMMC algorithm in data accuracy, so the AMLP algorithm is adaptive to the sensor network applications which require high accuracy data, such as geological hazard monitoring, medical care and military standard requirement.
出处
《传感技术学报》
CAS
CSCD
北大核心
2013年第4期550-557,共8页
Chinese Journal of Sensors and Actuators
基金
国家自然科学基金项目(21076179)
浙江省科技厅科技计划项目(2012C31014)
