As research on large regions of earth progresses, many geographical subdivision grids have been established for various spatial applications by different industries and disciplines. However, there is no clear relationship between the different grids and no consistent spatial reference grid that allows for information exchange and comprehensive application. Sharing and exchange of data across departments and applications are still at a bottleneck. It would represent a significant step forward to build a new grid model that is inclusive of or compatible with most of the existing geodesic grids and that could support consolidation and exchange within existing data services. This study designs a new geographical coordinate global subdividing grid with one dimension integer coding on a 2ⁿ tree (GeoSOT) that has 2ⁿ coordinate subdivision characteristics (global longitude and latitude subdivision) and can form integer hierarchies at degree, minute, and second levels. This grid has the multi-dimensional quadtree hierarchical characteristics of a digital earth grid, but also provides good consistency with applied grids, such as those used in mapping, meteorology, oceanography and national geographical, and three-dimensional digital earth grids. No other existing grid codes possess these characteristics.

The role of on-orbit computing for satellites is transitioning from being a backup measure to becoming a primary key function. However, the limited computing resources available on satellites make it difficult to deploy advanced models with large parameters. Additionally, satellite on-orbit computing requires high speed and accuracy, posing significant challenges for developing suitable models. To overcome these challenges, we propose an efficient vision transformer, RepSViT, for satellite on-orbit computing. The RepSViT introduces Spiking neural networks (SNNs) with high biological plausibility, event-driven property and low power consumption into the field of remote sensing image processing and satellite on-orbit computing for the first time and incorporates structural reparameterization. Specifically, we design a dynamic dilated spiking convolution (D ² SC) based on SNNs to improve the feature extraction capability and efficiency of RepSViT. We also develop a spiking guided attention module (SGAM) to make RepSViT pay more attention to object-related features with lower computational costs. Furthermore, we design an efficient coupled fine–coarse-grained block (ECFC) to enhance the model’s capability in extracting coarse and fine-grained features. To ensure effective feature extraction, inference speed and reduced computational costs, we design a reparameterized feed-forward network (RepFFN). RepSViT achieves an inference latency of 8.33 ms and a recognition accuracy of 95% on an embedded GPU, utilizing 3.77 million parameters and consuming 0.6 GFLOPs computational costs.

Forecasting the motion of surrounding vehicles is necessary for an autonomous driving system applied in complex traffic. Trajectory prediction helps vehicles make more sensible decisions, which provides vehicles with foresight. However, traditional models consider the trajectory prediction as a simple sequence prediction task. The ignorance of inter-vehicle interaction and environment influence degrades these models in real-world datasets. To address this issue, we propose a novel Dynamic and Static Context-aware Attention Network named DSCAN in this paper. The DSCAN utilizes an attention mechanism to dynamically decide which surrounding vehicles are more important at the moment. We also equip the DSCAN with a constraint network to consider the static environment information. We conducted a series of experiments on a real-world dataset, and the experimental results demonstrated the effectiveness of our model. Moreover, the present study suggests that the attention mechanism and static constraints enhance the prediction results.

Flight conflict detection is fundamental to flight dispatch, trajectory planning, and flight safety control. An ever-increasing aircraft population and higher speeds, particularly the emergence of hypersonic/supersonic aircrafts, are challenging the timeliness and accuracy of flight conflict detection. Traditional trajectory conflict detection algorithms rely on traversing multivariate equations of every two trajectories, in order to yield the conflict result and involve extensive computation and high algorithmic complexity; these algorithms are often unable to provide the flight conflict solutions required quickly enough. In this paper, we present a novel, low-altitude flight conflict detection algorithm, based on the multi-level grid spatiotemporal index, that transforms the traditional trajectory-traversing multivariate conflict computation into a grid conflict state query of distributed grid databases. Essentially, this is a method of exchanging "storage space" for "computational time". First, we build the spatiotemporal subdivision and encoding model based on the airspace. The model describes the geometries of the trajectories, low-altitude obstacles, or dangerous fields and identifies the grid with grid codes. Next, we design a database table structure of the grid and create a grid database. Finally, we establish a multilevel grid spatiotemporal index, design a query optimization scheme, and examine the flight conflict detection results from the grid database. Experimental verification confirms that the computation efficiency of our algorithm is one order of magnitude higher than those of traditional methods. Our algorithm can perform real-time (dynamic/static) conflict detection on both individual aircraft and aircraft flying in formation with more efficient trajectory planning and airspace utilization.

The Shanghai New Year’s Eve stampede on 31 December 2014, caused 36 deaths and 47 other injuries, generating attention from around the world. This research aims to explore crowd aggregation from the perspective of Sina Weibo check-in data and evaluate the potential of crowd detection based on social media data. We develop a framework using Weibo check-in data in three dimensions: the aggregation level of check-in data, the topic changes in posts and the sentiment fluctuations of citizens. The results show that the numbers of check-ins in all of Shanghai on New Years’ Eve is twice that of other days and that Moran’s I reaches a peak on this date, implying a spatial autocorrelation mode. Additionally, the results of topic modeling indicate that 72.4% of the posts were related to the stampede, reflecting public attitudes and views on this incident from multiple angles. Moreover, sentiment analysis based on Weibo posts illustrates that the proportion of negative posts increased both when the stampede occurred (40.95%) and a few hours afterwards (44.33%). This study demonstrates the potential of using geotagged social media data to analyze population spatiotemporal activities, especially in emergencies.

With the enormous scale of remote sensing imagery generation, on-orbit computing has become a crucial paradigm to enable near real-time processing. Due to the limited onboard resources and on-orbit power supply, satellite edge computing (SEC) is developed for satellite-ground collaboration, aiding on-orbit computation. However, the intermittent satellite-to-ground transmission link poses an efficiency challenge when collaborating SEC resources. Therefore, this paper proposes a Satellite Edge Computing resource scheduling technique for On-orbit Remote Sensing imagery processing (SECORS). First, we design a remote sensing mission-specific SEC architecture, which involves an offline-online satellite working mode. Subsequently, a computational resource scheduling model (SEC-RSM) is established, including the directed acyclic graph (DAG) model and mathematical problem formulation. Next, to obtain effective scheduling solutions, we develop an end-to-end algorithm leveraging the multi-agent proximal policy optimization and heuristic rule of the earliest finish time (SEC-MPH). Finally, we build a simulation SEC platform to carry out experiments and implement several methods as the comparison including multi-objective evolutionary algorithms, deep reinforcement learning approaches, and the scheme without optimization (baseline). Simulation results show that SECORS achieves 68.87% and 66.60% reductions in time and energy for on-orbit computation. Moreover, our method improves the energy efficiency ratio by three times and achieves high processing capacity with 548 pixels per unit of power (W) and time (ms).