信号场强压缩感知的传感器定位方法研究

发布时间：2018-06-16 20:57

本文选题：信号场强叠加 + 压缩感知　；参考：《仪器仪表学报》2014年06期

【摘要】：提出了多包接收时的信号场强叠加模型,建立了观测场强与传感器位置的映射关系。由于传感器数量相对于网格数量是稀疏的,将传感器定位转化为压缩感知问题求解,以减少观测的信号数量,并提出了NL1-norm算法计算出传感器的位置。通过数值仿真,分析了传感器信号功率、观测信号数量以及传感器个数对定位误差的影响。相同条件下,验证了NL1-norm算法的定位精度相比最小化L1-norm算法和贪婪匹配追踪(GMP)算法提高了2倍。低信噪比情况下比较得出,基于CS的节点定位方法误差和观测代价都明显小于RSSI和MDS-MAP方法。
[Abstract]:The superposition model of the signal field strength is proposed and the mapping relationship between the observed field strength and the position of the sensor is established. Since the number of sensors is sparse relative to the number of meshes, the sensor location is transformed into a compressed sensing problem to reduce the number of observed signals, and the NL1-norm algorithm is proposed to calculate the position of the sensor. Through numerical simulation, the effects of sensor signal power, number of observed signals and number of sensors on positioning error are analyzed. Under the same conditions, the accuracy of NL1-norm algorithm is 2 times higher than that of minimized L1-norm algorithm and greedy matching tracking GMPalgorithm. In the case of low signal-to-noise ratio (SNR), it is found that the error and the observation cost of the CS-based node localization method are obviously lower than those of the RSSI and MDS-MAP methods.
【作者单位】：合肥工业大学计算机与信息学院;教育部安全关键工业测控工程研究中心;
【基金】：国家自然科学基金(61370088) 国家国际科技合作专项项目(20140FB10060)资助
【分类号】：TN929.5;TP212.9

【参考文献】