一种基于稀疏重构的NB-IoT时延估计算法

针对传统的时延估计算法无法解决窄带物联网(narrowband internet of things， NB-IoT)低速率低功耗引起的估计精度低、计算复杂度高等问题，提出一种加入代价函数模型的基于稀疏重构的时延估计算法。利用窄带定位参考信号(narrowband positioning reference signal， NPRS)与传统的正交匹配追踪算法(orthogonal matching pursuit，OMP)进行时延值预估计，然后根据预估计的时延值构建冗余字典，在此基础上利用改进的OMP算法进一步对时延值进行估计。该算法中加入代价函数的思想将多维的时延估计降维成多个一维的时延估计，同时利用稀疏重构来消除各信号之间的干扰。另外，为了消除降维带来的局部最优的问题，合理设置软门限来有效并快速地终止代价函数模型的迭代过程。仿真结果表明，与OMP算法等传统时延估计算法相比，该算法具有更好的检测性能以及更高的时延估计精度。

A time delay estimation algorithm of narrow band internet of things based on sparse reconstruction

Aiming at the problems of low estimation accuracy and high computational complexity caused by low rate and low-power consumption of Narrowband Internet of Things (NB-IoT), a new time delay estimation algorithm based on sparse reconstruction with cost function model is proposed. Firstly, the traditional OMP algorithm and narrowband positioning reference signal are used to predict the delay value, and then the redundant dictionary is constructed based on the pre-estimated delay value. Based on this, the improved OMP algorithm is used to further estimate the delay value. The idea of adding a cost function to the algorithm reduces the multidimensional delay estimation into multiple one-dimensional time delay estimation, and uses sparse reconstruction to eliminate the interference between signals. In addition, in order to eliminate the problem of local optimization caused by dimensionality reduction, the soft threshold is reasonably set to effectively and quickly terminate the iterative process of the cost function model. The simulation results show that compared with the traditional correlation algorithm, OMP algorithm and other delay estimation algorithms, the algorithm has better detection performance and higher delay estimation accuracy.