Radiation belts have been observed at many magnetized planets in the solar system, such as Earth (Summers et al., 2009), Jupiter (Bolton et al., 2004) and Saturn (Carbary et al., 2009. Understanding the transport, acceleration and loss of radiation belt electrons is a major research topic in planetary magnetospheric physics (Tang & Summers, 2012;Tang et al., 2014) since these energetic electrons pose a potential hazard to orbiting satellites. Previously, the observational data being gathered by numerous satellites were incorporated into simulation models of the radiation belt (Ni et al., 2019), which remain widely used for spacecraft design purposes and space environment predictions. However, the practical space weather applications also require innovative approaches to the use of mass satellite data, and will lead to further advances in the research of planetary radiation belts (Ni et al., 2009).With the continuous development of artificial intelligence technology, scientists began to explore the applications of machine learning in space physics, such as the stellar atmospheric parameter (Ramirez et al., 2001),

With the rapid progress of hardware and software, a wireless sensor network has been widely used in many applications in various fields. However, most discussions for the WSN node deployment mainly concentrated on the two-dimensional plane. In such a case, some large scale applications, such as information detection in deep space or deep sea, will require a good three dimensional (3D) sensor deployment scenario and also attract most scientists’ interests. Excellent deployment algorithms enable sensors to be quickly deployed in designated areas with the help of unmanned aerial vehicles (UAVs). In this paper, for the first time, we present a three dimensional network deployment algorithm inspired by physical dusty plasma crystallization theory in large-scale WSN applications. Four kinds of performance evaluation methods in 3D space, such as the moving distance, the spatial distribution diversion, system coverage rate, and the system utilization are introduced and have been carefully tested.Furthermore, in order to improve the performance of the final deployment, we integrated the system coverage rate and the system utilization to analyze the parameter effects of the Debye length and the node sensing radius. This criterion attempts to find the optimal sensing radius with a fixed Debye length to maximize the sensing range of the sensor network while reducing the system redundancy. The results suggest that our 3D algorithm can quickly complete an overall 3D network deployment and then dynamically adjust parameters to achieve a better distribution. In practical applications, engineers may choose appropriate parameters based on the sensor’s hardware capabilities to achieve a better 3D sensor network deployment. It may be significantly used in some large-scale 3D WSN applications in the near future.

Internet of Things (IoT) and Big Data technologies are becoming increasingly significant parts of national defense and the military, as well as in the civilian usage. The proper deployment of large-scale wireless sensor network (WSN) provides the foundation for these advanced technologies. Based on the Fruchterman–Reingold graph layout, we propose the Fruchterman–Reingold Hexagon (FR-HEX) algorithm for the deployment of WSNs. By allocating edges of hexagonal topology to sensor nodes, the network forms hexagonal network topology. A comprehensive evaluation of 50 simulations is conducted, which utilizes three evaluation metrics: average moving distance, pair correlation diversion (PCD), and system coverage rate. The FR-HEX algorithm performs consistently, the WSN topologies are properly regulated, the PCD values are below 0.05, and the WSN system coverage rate reaches 94%. Simulations involving obstacles and failed nodes are carried out to explore the practical applicability of the FR-HEX algorithm. In general, the FR-HEX algorithm can take full advantage of sensors’ hardware capabilities in the deployment. It may be a viable option for some IoT and Big Data applications in the near future.