Three-dimensional ionospheric tomography based on compressed sensing

Abstract The insufficient number of low-elevation observations is a limitation the three-dimensional ionospheric computer tomography (CT) based on global navigation satellite system (GNSS). To solve this problem, accurate prior information regional ionosphere must be obtained. However, it difficult to explicitly and accurately express information. This study uses compressed sensing (CS) for first time. Specifically, electron density obtained from international reference used build dictionary fully integrate into dictionary. Then, reconstructed by using compressive sampling matching pursuit method. Subsequently, GNSS data China (Region I) Europe II) were utilized validate proposed method, results are compared with ionosonde observations. mean standard deviation (SD) difference respect result 41 22 km, respectively. SD relative 16% 9%, In Region II, between reversed peak 1.9 × 10 m ?3 8.1 , 3% 13%, height 33 19 11% 7%. distribution variation in these two regions showed local time dependence, horizontal gradient latitude was greater than that longitude. Moreover, CT CS efficient, taking about 6 s per inversion an desktop 16 GB RAM Intel (R) Core (TM) i7-8700 CPU.