Tohoku University Achieves Daily Management of Eight Satellites with New System

Researchers are increasingly reliant on Earth observation satellites for vital data, yet managing multiple missions and ground stations presents significant logistical challenges. Yuji Sakamoto from Green Goals Initiative and Tohoku University, alongside colleagues, detail a novel on-demand satellite operation system developed and refined since 2009, demonstrating efficient management of up to eight operational satellites simultaneously. Their work, presented in this report, showcases how cloud-based functions and automatic command generation , utilising ground stations in Sendai, Hakodate and Sweden , can streamline satellite operations and represents a substantial step towards scalable and responsive space-based Earth observation.

The research, stemming from a continuous development program since the launch of their first satellite in 2009, details the operational achievements and introduces the innovative system underpinning this enhanced capability. This system allows for streamlined Earth observation missions and engineering demonstration projects utilising both 50cm-class satellites and CubeSats up to 3U in size.

The team developed a sophisticated system capable of collecting approximately 1 GB of image data per pass via a 20 Mbps X-band link, demonstrating a substantial increase in data throughput. This was achieved through the integration of cloud-based functions for both satellite control and ground station operations, significantly reducing the time and effort required for mission planning and execution. Crucially, the system facilitates “on-demand satellite operation” by supporting multiple satellites, ground stations, and users, even those outside the traditional space sector, with a scalable database server capacity. The core of this web service is the Satellite Operation Management (SOM) system, which orchestrates communication opportunities between ground stations and satellites, manages imaging requests, and distributes data online.
Experiments show that the system dramatically reduces operational planning time, achieving up to a 90% reduction in preparation efforts. This is accomplished by automatically generating time lists for communication and imaging opportunities, and by employing predefined command templates that are dynamically adjusted for specific tasks. While complete automation remains a challenge due to satellite-specific variations, the system allows operation planners to focus on critical final checks, such as power and memory management, thereby minimising human error. Detailed explanations of the SOM and the associated Satellite Data Management (SDM) function are presented, alongside a comprehensive history of 13 satellites managed as of May 2025, including those operated by affiliated groups.

The research highlights the successful operation of several key satellites, including RISING-2, which achieved 5-meter resolution multi-wavelength imaging, and DIWATA-2, equipped with the SMI+ERC sensor for 54-meter resolution multi-wavelength imaging and the HPT camera. Despite an early malfunction with SPRITE-SAT, signal monitoring continued for 15 years, demonstrating the robustness of the ground station infrastructure. Currently, DIWATA-2 and HIBARI remain operational, with plans to integrate three new observation targets in 2025, 2026. The team developed a cloud-based system enabling automated command generation and streamlined operations, significantly reducing planning time for complex missions. Researchers pioneered a system that automatically calculates optimal imaging times and communication windows, generating command sequences tailored for each satellite.

This innovative approach incorporates predefined templates for routine tasks, automatically correcting execution times and target points, although complete automation remains challenging due to satellite-specific planning variations. Crucially, the system reduces preparation time by up to 90%, allowing operation planners to focus on final, critical checks of satellite settings like power and memory allocation, thereby minimising human error. The study details the Satellite Operation Management (SOM) and associated Satellite Data Management (SDM) functions, tracking a total of 13 satellites managed as of May 2025, categorised by size, 50cm microsatellites (M), 2U CubeSats, and 3U CubeSats. Experiments employed a tiered ground station infrastructure, with CRESST equipped for U-band command, S-band command, S-band telemetry (up to 100 kbps), and high-speed X-band telemetry (up to 20 Mbps).

HU-Kiruna supports S-band telemetry (up to 1 Mbps), while HU-Hakodate provides X-band telemetry (up to 20 Mbps), enabling diverse data acquisition capabilities. RISING-2, the first satellite to deliver meaningful multi-wavelength images, achieved 5-meter ground resolution using a U-band command and 100 kbps S-band telemetry system. Subsequent satellites, like DIWATA-2, integrated advanced sensor systems, the SMI+ERC providing 54-meter multi-wavelength imaging and the HPT camera for high-resolution imaging. The addition of X-band telemetry, peaking at 2.4 Mbps, significantly increased data download volumes, a trend continued with RISING-4’s 20 Mbps X-band link and 47-meter resolution SMI sensor.

Data analysis reveals that RISING-4 achieved the highest cumulative downlink volume at 171 GB over its 2 years and 1 month operational lifespan, with CRESST recording the largest volume at 131 GB. HU-Kiruna, leveraging its polar location and link margin, also achieved substantial data volumes despite utilising S-band telemetry, demonstrating the effectiveness of the combined ground station network0.1km, and the area of Nishinoshima at 4. The research details operational achievements and introduces the supporting system enabling this efficient satellite management. The core of this achievement lies in cloud-based management functions, which facilitate both satellite and ground station operations through automatic command generation, a process refined through initial operational phases and template-based parameter replacement.

This system streamlines tasks like setting observation times and target coordinates, though the current implementation relies on complex conditional branching resembling source code, presenting an area for future optimisation. Manual operator intervention is still required when scheduling overlapping communication passes or observation targets, indicating a need for further automation. Acknowledging these limitations, the authors highlight the potential for system refinement through continued learning from practical satellite operations. Future work will focus on enhancing automation capabilities and improving the efficiency of command generation processes. These advancements promise to further optimise the management of small satellite constellations, contributing to more effective Earth observation and space-based engineering demonstrations.