8 Nov 2025: It’s been a while since I checked in on the Russian Cosmos 2581/82/83 (62902/62903/62904) trio. Recall Russia launched these three satellites on 5 Feb 2025 (video) and released no details of the mission other than to say they were for the “Ministry of Defense.” While my reporting regarding these objects ceased in early April, their activities have not…the spacecraft regularly orbit within <1km of one another. My ability to analyze what’s going on is limited due to reliance on two-line element set (TLE) based data which is satisfactory for monitoring the general goings-on on orbit but presents challenges when trying to perform detailed analysis of two objects which are extremely close to one another or frequently maneuvering (or both). With these limitations I turned to the pros over at COMSPOC_OPS to take a look at the trio and compare our results. To limit the scope of the research we focused on on just the interactions from 24-29 October 2025. Watch Video of 25 Oct Activity. – First a little review:

• Russia has maneuvered Cosmos 2581 the least going back to Feb 2025. It seems Cosmos 2581 plays the role of “target” for both Cosmos 2582 and 2583 which maneuver much more frequently.
• In Feb Russia maneuvered Cosmos 2581 & 2582 into the same orbital altitude and initiated a long series of proximity maneuvers. They remained in close vicinity from Feb – Apr 2025.
  • During this period, 2581 & 2582 achieved extremely close distances. COMSPOC reported they came as close as 50m from one another while flying in tandem. On 11 Mar one report indicated a close approach to be within 15 meters.
  • During this time Cosmos 2582 was the active satellite performing the maneuvering, with Cosmos 2581 acting as the target.
  • For the details see 9 Mar 2025 & 6 Apr 2025 Flash and associated video.
• During this time Cosmos 2583 did not participate in proximity operations. However, it did release a sub-satellite (Obj F) on 18 Mar and did conduct proximity operations with it through May 2025.
  • This may have been a test of Cosmos 2583 systems. In late May Russia raised Cosmos 2583’s orbit and in mid-June closely approached Cosmos 2581, coming within a few hundred meters or less. Subsequently, Cosmos 2582 joined them by maneuvering on June 23, resulting in all three satellites orbiting closely together.
• During the period of our research (24-29 Oct 2025) Cosmos 2583 conducted several close approaches with Cosmos 2581. Cosmos 2582 played no apparent role in the operations as it remained over 250km away and is orbiting at a slightly higher orbit than its sibling satellites.

– TLE Based Analysis 24-29 October 2025

• Cosmos 2581 and Cosmos 2583 are in RPO conditions with one another. For the majority of the time the two satellites were within 10km of one another. There were 3 distinct periods where the satellites were within 1km of one another. The point of closest approach was ~463m. Both satellites have an average altitude of 585.3 to 585.0km.
• Cosmos 2582 was not involved with this particular set of close approaches, its point of closest approach with either 2581 or 2583 was over 250km. The average altitude of Cosmos 2582 was 585.3km during this time period.

– COMSPOC Analysis: Cosmos 2581 v 2583 24-29 October 2025

• TLE analysis provides some indications of interesting activity, but lacks fidelity to gain an understanding of what is actually happening. TLEs consist of average results from multiple observations which is helpful when monitoring/cataloging space objects which are maneuvering infrequently or not operating in close proximity with other satellites. To understand what was happening between Cosmos 2581 and 2583 we needed more accurate information. Thankfully the COMSPOC team has access to highly accurate data.
• One of the issues when using TLEs to monitor object orbiting in close proximity is called “cross tagging” which occurs when one object is misidentified with another resulting in faulty orbital information. After noticing an increase in cross tags between COSMOS 2581 and 2583, COMSPOC’s ops team scrutinized the behavior between these objects using data from LEOLABS and their COMSPOC’s kalman filter-based solution, produced from one of

their applications called “SSASuite.”

• Reviewing raw sensor data enabled COMSPOC orbital analysts to resolve several cross tag instances. Use of the superior precision of the High Definition ephemeris, enabled by LeoLabs data, proved crucial to the task and allowed a level of analysis not possible using only TLEs.
• Point of Closest Approach: On 25 Oct 2025 10:01:41z COSMOS 2583 approached within 80 meters of COSMOS 2581, passing from below to above the VVLH X-Y plane of COSMOS 2581 (see graphic & video).

Russian space operators likely set up this point of closest approach with 2 separate Cosmos 2583 maneuvers. The first occurred on between 1710-0744z with a total delta v (fuel usage, “dV”) of 0.242 m/s, and the second just over an hour later between 0851-0903z with a total dV of 0.407 m/s. You can see both maneuvers in this VIDEO, specifically when Cosmos 2583 is highlighted in RED.

• COMSPOC analysts also produced two other videos (vid1 & vid2) that show the close approach window from different perspectives.

– In some of the videos you’ll notice the satellites are in the center of an (American) football shaped cage. These cages are covariance ellipsoids which indicate COMSPOC’s confidence in the solution. A covariance ellipsoid, such as you see in the video, represents a volume within which COMSPOC’s system has determined that the object resides, with a 95% certainty. The radii of the covariance volume for COSMOS 2583 are 221.1m by 22.92m by 52.95m. By contrast, the covariance radii for COSMOS 2581 are 61.58m by 14.75m by 10.92m. This volume is considerably smaller than 2583’s, indicating that COMSPOC had better precision about the position of 2581 than 2583. And given that 2583 was maneuvering during this time, it makes clear sense why this would be.

• • Editor’s comment: Russia is clearly testing RPO capabilities and tactics with Cosmos 2581, 2582 and 2583. In our 5-day focus there were examples of 2581 and 2583 coming within <100m of one another. It will be interesting to see if Cosmos 2582 again enters the fray. Special thanks to the COMSPOC team for their support!