Project Overview
As satellite launches accelerate, over 46,000 trackable space objects ("The US Space Force - America's Invisible Front Line," May 1, 2025) and over 500,000 smaller space debris threaten collisions with satellites. Nations with partial Space Situational Awareness (SSA) and unknown SSA capabilities rely primarily on the United States' systems, which unfortunately, yields errors for non-U.S. satellites, no real collision warnings, and incomplete coverage, leaving them without reliable warnings. These countries need an independent, yet interoperable system that overcomes their current challenges. Our solution aims to provide a software as a service platform for these countries that allow them to cooperatively contribute, share and access orbital data, as well as obtain anomaly alerts through an integrated anomaly detection engine. Additionally, the platform is intented to function under the stewardship of the United Nations Office for Outer Space Affairs (UNOOSA) and the International Telecommunication Union (ITU), in order to establish the diplomatic, technical, and regulatory infrastructure needed for neutral governance.
Business Understanding
Space is becoming increasingly more crowded. In the last 60 years nearly 8,500 objects have been launched to space, about 1,500 in geosynchronous Earth orbit (GEO) and about 7,000 in low Earth orbit (LEO)(Oltrogge & Alfano, 2019). A large fraction remains especially in LEO. Going forward, LEO is expected to become more crowded. The concern is not just the increasing number of satellites and active payloads, but the amount of debris (rocket bodies, other inert bodies, dead payloads) in Earth’s orbit, which comprises more than 95 percent of the currently tracked objects in space.
It is not just the number of satellites that is increasing—the number of satellite operators has been increasing steadily over the last 60 years. More countries have become active in space. Smaller, lighter, and more capable satellites make Earth observation and remote sensing within the reach, not just of countries, but also corporations and individuals. For example, Bank Rakyat in Indonesia launched satellite BRIsat in 2016, built by Space System/Loral and launched by Arianespace, to manage its 50 million accounts. Another example is the CubeSat satellite launched by NanoRacks in 2019 under the auspices of Zac Manchester who was able to afford it through crowdfunding. This trend is likely to continue, making the space environment increasingly more crowded (Oltrogge & Alfano, 2019).
Although relatively few catastrophic collisions have occurred thus far in space, the likelihood of a collision is predicted to increase in the near future, given the expected growing number of objects in both LEO and GEO and limited ability to track objects’ orbits, which will make it difficult for operators to adequately avoid threats(Oltrogge & Alfano, 2024). This problem may be exacerbated if any of the proposed constellations of small satellites in LEO are launched, as they will dramatically increase the number of objects that require tracking, thus increasing the tracking and computational requirements for SSA in general, and conjunction warnings in particular.
Given that the systems developed to track space objects were developed at a time when there were fewer objects in space, the accuracy of prediction is low. Oftentimes, a DoD conjunction warning message has an error ellipse of 100 km or more(Oltrogge & Alfano, 2024); and the rate of false positives is high as well ((Oltrogge & Alfano, 2024). Because of these two factors, as traffic in space grows, both the number of conjunction warning messages, as well as the rates of both false positives—and false negatives— are likely to increase.
As space capabilities become integral to more applications (e.g., earth observation, communications, global positioning), a growing number of countries are recognizing the security and economic value of space, and consequently, increasing their spending in space. The number of countries involved in space continues to increase. A decade ago, fewer than 50 countries were investing in space. Today, there are 70 (BIS Research, 2022). In the coming decade, that number is expected to increase to more than 80 countries (BIS Research, 2022).
Many stakeholders indicated that they need to have trust and confidence in the data being shared for collision warnings and other SSA products (Lal et al., 2018). Many acknowledged the usefulness of verifying the information that is part of any database(Lal et al., 2018). There are many concerns with the current systems for provision of SSA. Some operators question the accuracy and especially the completeness of the information provided to them by the DoD (Lal et al., 2018). For example, some South Korean government officials estimate that their country receives data on only about 40 percent of the objects tracked by the DoD, due to sensitivity of U.S. assets (Oltrogge and Alfano, 2019). This distrust is further complicated by the lack of transparency related to computing outcomes such as probabilities of collision (Lal et al., 2018). Owners and operators believe they require more information of high enough quality to make well-informed decisions about maneuvering. But because they do not know the process by which U.S.-provided information on an object’s location is processed into a collision assessment or warning, they often do not feel confident maneuvering based on that warning (Oltrogge and Alfano, 2019).
Given the countries’ dependence on space for critical national needs and societally critical endeavors, SSA is becoming important enough that most countries active in space activities want to establish SSA systems that are more independent in tracking objects in space (Lal et al., 2018).
Some nations are looking to decrease their dependence on the U.S. SSA system due to concerns that it might be the target of an adversarial attack(Lal et al., 2018). Other stakeholders felt such an attack is unlikely, given the global nature of SSA and the widespread reliance on the U.S. system. Some (e.g., France) seek to be self-sufficient for reasons of national pride and sovereignty (Lab, 2020). Others see it as a means by which to provide leadership in the domain and collaborate with 23 other nations (Lal et al., 2018). Many countries that are pursuing their own SSA systems, explicitly intend to keep their systems interoperable with others internationally (e.g., Australia, Canada, Japan, United Kingdom) (Lab, 2020).
As the number of satellites and debris in orbit continues to grow, no single country can track and manage space traffic effectively on its own. A SaaS Global Space Traffic Management (STM) platform offers a mechanism where countries can voluntarily contribute orbital and object data to a global system, but do so in a way that protects their independence and national priorities. By using standardized formats such as TLEs (Two-Line Elements), OMMs (Orbit Mean Elements Messages), and CDMs (Conjunction Data Messages), each nation can share only the subset of its data that is approved for publication. This ensures that sensitive or classified information remains under national control while still allowing the broader international community to benefit from a more transparent, aggregated picture of orbital activity.
For participating countries, interoperability is achieved by adhering to agreed-upon international data standards and protocols. Nations such as Australia, Canada, Japan, and the United Kingdom have already expressed intentions to build independent SSA systems that are nevertheless interoperable with international frameworks (Lal et al., 2018). This model allows countries to preserve their sovereignty—retaining their own tracking networks, analytical methods, and decision-making authority—while ensuring their data can “plug into” a shared global platform. In practice, this means each government could run its own SSA system domestically but also choose to upload selected orbital information, such as satellite registries or debris tracking data, into the SaaS platform. Once uploaded, the data is normalized and displayed alongside contributions from other countries, increasing the SSA’s capabilities in identifying potential collisions, congestion zones, or long-term sustainability risks in a way that no single nation could achieve alone(Airbus Ventures, 2019).
Crucially, the SaaS platform supports tiered data-sharing agreements, so each participant grants viewing permissions and alerts according to its requirements. For instance, a country may decide to make its active satellite catalog publicly visible, share conjunction alerts with trusted partners only, and keep precise maneuver plans restricted to internal use. This approach addresses the concerns of achieving greater transparency and increased confidence in collision predictions, while protecting national security(Airbus Ventures, 2019).
Benefits of the System:
- By letting countries decide their level of disclosure, the system lowers political barriers to participation while increasing the overall volume and quality of global orbital data made publicly available.
- Participation provides tangible benefits to each country. By contributing to the system, governments gain access to higher-quality conjunction warnings, better situational awareness of crowded orbits, and the ability to verify or cross-check information provided by others. This reduces reliance on a single foreign system, which is a growing concern for countries worried about over-dependence on U.S. data.
- Participation builds international trust and transparency, as all parties can see how their data contributes to a collective orbital “map.”
- Countries retain their independence, safeguard their national interests, and strengthen their own SSA capabilities, as they benefit from the efficiencies and global reach of a cooperative SaaS-based STM platform.
Data Understanding
Our model's data source is Space-Track.org, extracted on August 26, 2025 at 12:45 a.m. East African Time. Our query parameters were Country = PRC, Object_Name, NORAD_CAT_ID, INTLDES, LAUNCH_DATE, and TLE_Data. A data dictionary is included in this repository for column name explanations. Lastly, the query returned 8,648 records entered since 1970. This dataset is in the file named satellite_data.json.
Modeling and Evaluation
Two models were used to cluster the dataset (DBSCAN) and identify anomalies (IsolationForest). DBSCAN is useful because it doesn’t need the number of clusters specified in advance, and can handle noise. The DBSCAN performance metrics below demonstrates that we achieved strong results. Also, the metrics confirm that DBSCAN created well-formed clusters that matched Earth's orbital regimes--low earth orbit (LEO), medium earth orbit (MEO), and geostationary earth orbit (GEO). These scores are:
- Silhouette Score: 0.835 (good separation, close to 1)
- Davies-Bouldin Index: 0.140 (very low, good compactness/separation)
- Calinski-Harabasz Index: 10738 (high, strong clustering quality)
Lastly, the cleaned dataset returned cluster counts as follows:
- Cluster 0 → 8073 satellites
- Cluster 1 → 147 satellites
- Cluster 2 → 258 satellites
- Cluster 3 → 64 satellites
- Noise (-1) → 90 satellites
Then we applied IsolationForest to each non-noise cluster, in order to flag unusual satellites. By combining the object type from the object name data with the anomaly output from IsolationForest, we were able to infer the following anomalies:
- Debris drifting away from its group
- Rocket bodies (R/B), possibly tumbling or decaying
- Active satellites with motions not matching their cluster neighbors
This dual DBSCAN and IsolationForest approach connects machine learning with real orbital science, such that the clusters are not just math outputs. Rather, they are meaningful categories that fall into LEO/MEO/GEO classifications. Additionally, the flagged debris, rocket bodies, and unusual satellites were detected anomalies. Consequently, our model, with its strong performance metrics, provides both space traffic insights (through clustering) and an anomaly detection engine that is data-driven and grounded in physics.
Conclusion
Although a prototype at this stage, our Orbit Guardians Global SaaS platform transforms global space safety. It enables voluntary data sharing, accelerates access to accurate orbital intelligence, and delivers collaborative anomaly detection for all subscribing nations. Most importantly, it provides affordable SSA access for developing SSA countries through neutral UNOOSA stewardship, eliminating dependence on a single nation's control, and making the future of space collaborative, sustainable, and accessible to all.
References:
Airbus Ventures. (2019, August 20). Z-SPACE: LeoLabs sets its sights on space situational awareness and space debris. Medium. https://medium.com/@airbus_ventures/z-space-leolabs-sets-its-sights-on-space-situational-awareness-and-space-debris-83004efb7d79
BIS Research. (2022, August 5). Key Technologies Enabling Enhanced Space Situational Awareness Sensors. Medium. https://bisresearchreports.medium.com/key-technologies-enabling-enhanced-space-situational-awareness-sensors-983be0f25db4
Lab, M. M. (2020, February 12). Space Situational Awareness: Key issues in an evolving landscape | Statement of Dr. Danielle Wood. Medium; MIT MEDIA LAB. https://medium.com/mit-media-lab/space-situational-awareness-key-issues-in-an-evolving-landscape-statement-of-dr-danielle-wood-9085d09d4b13
Lal, B., Balakrishnan, A., Caldwell, B. M., Buenconsejo, R. S., & Carioscia, S. A. (2018, April). Global Trends in Space Situational Awareness (SSA) and Space Traffic Management (STM). Dtic.mil. https://apps.dtic.mil/sti/html/trecms/AD1123106/
Oltrogge, D. L., & Alfano, S. (2019). The technical challenges of better Space Situational Awareness and Space Traffic Management. Journal of Space Safety Engineering, 6(2), 72–79. https://doi.org/10.1016/j.jsse.2019.05.004
Oltrogge, D., & Alfano, S. (2024). Innovative launch collision avoidance (LCOLA) tool prioritizing accuracy, launch access and efficiency. Journal of Space Safety Engineering, 11(2), 243–251. https://doi.org/10.1016/j.jsse.2024.04.011
United States Space Force (2025). The US Space Force - America's Invisible Front Line (A Documentary Film). https://www.youtube.com/watch?v=Qn938Gm6axQ