仅依赖连通度的压缩感知多目标定位方法

提出仅依赖连通度的多目标定位方法，将多目标定位问题转化为基于压缩感知的稀疏向量重构，解决室内参照物高密度分布的目标定位问题。定位方法仅以连通度为观测值，运用最小化l1-范数法求解目标位置。当观测数据压缩为1 bit时，提出半正定松弛和不动点迭代法结合的目标求解算法。根据仿真实验结果，与MDS-MAP、DV-Hop和RSS-CS方法进行比较得出，仅连通度的非1-bit和1-bit量化的CS定位方法的平均定位误差小于1个网格，且2种方法占用的比特数只相当于RSS定位方法占用比特数的 和 。

Multiple targets localization via compressive sensing from mere connectivity

A multiple targets localization method was proposed from mere connectivity, and the multiple targets positioning problem was converted to sparse vector resolving by compressive sensing theory, which was applied to the indoor localization of intensive references distribution. The connectivity to the references was collected as the only measurement data, and targets locations were figured out by minimum l1-norm algorithm. When measurement data was compressed to 1 bit, the fixed point iteration algorithm combined with semi-definite relax was proposed to figure out targets locations. As for the simulation re-sults, compared with MDS-MAP, DV-Hop and RSS-CS algorithms, the average location error is less than 1 grid by the mere connectivity of N-bit and 1-bit quantization CS localization, of which the occupied bit quantity are reduced to less than 4 times and 16 times of RSS localization observation value respectively.