In an increasingly crowded orbital environment, collision avoidance is no longer a manual task, and relying on partial or siloed data is becoming an unacceptable risk. The next era of space operations must be built on shared maneuver data, high-fidelity ephemerides, smart data fusion and intelligent automation.
Read our top four takeaways from our conversation with Dr. Siamak Hesar, CEO of Kayhan Space, or listen to the full episode.
Takeaway 1: Siloed maneuver data is a threat to orbital safety.
As satellite numbers climb into the tens of thousands, Hesar warned that operators withholding maneuver plans poses a major systemic risk. “One of the big gaps is the fact that the operators do not share their planned maneuvers with each other by and large.”
Without that advance notice, even predicted non-events can turn into last-minute conjunction alerts, or worse.
“Imagine you’re driving on the road and you don’t turn on your turn signal and you just turn,” he said. “That turn signal is sharing planned future maneuver plans. No one else in the world other than that particular operator knows about the existence of that plan.”
Takeaway 2: Ephemeris sharing is the backbone of space traffic coordination.
Hesar described ephemeris as a time series of position and velocity of a satellite. This predicted orbital path is what enables coordinated avoidance planning. Without it, even the most sophisticated tracking systems can’t anticipate where a spacecraft will be hours or days into the future.
“If you don’t know where something is going to be, then you cannot coordinate,” he said.
Hesar also clarified that the most important aspect is simply the maneuver plan itself, not proprietary technical details.
“The amount of data that we require to make an informed decision is very minimal,” he said. “We’re not talking about propulsion technology, just the thrust vector that is going to result in changing of the trajectory.”
Takeaway 3: Data fusion and AI are transforming conjunction assessment.
With government sensors, commercial tracking networks, GPS telemetry and operator-provided data all offering different levels of precision, managing orbital awareness is increasingly a data-fusion challenge.
“We are working very hard to ensure we are able to pull data from different sources to have a better picture of what is happening in space,” Hesar said.
Kayhan uses multiple phenomenologies—radar, optical, GPS-based ephemerides—to offset the limitations of any single source. “Certain things might not be very visible with one particular type of data.”
On the role of AI, Hesar drew a sharp line between physics-based modeling and system-level optimization. However, AI becomes powerful in higher-level tasks such as anomaly detection.
“That’s a very good use case to teach them about what is the ‘normal’ behavior of an object and detect if there is an anomalous action,” he said.
Takeaway 4: Automation will enable orbital scalability.
Conjunction alerts already number in the tens of thousands per day. With the rapid rise of mega-constellations, manual collision avoidance will become impossible.
Hesar pointed to Starlink as proof that automation can handle complexity. “They have more than 8000 operational satellites. How was that done? Through automation,” he said.
But even the best autonomous systems fail when neighbors don’t share information.
“The problem is not that we don’t have the capability of operating a large number of satellites,” he said. “The problem comes back to lack of communication.”
Orbital safety is not a technology problem, but rather a coordination problem. What’s missing is universal adoption of shared maneuver data, consistent ephemeris sharing and automated coordination frameworks. But if operators collaborate, intelligent automation can make the orbital environment safer, more transparent and more resilient.
For more, listen to the full episode.
Explore More:
Falling Debris and Near-Miss Collisions: Mitigating Risk in LEO
Rockets, Re-use and Return-to-Earth: Novel Approaches to Orbital Access
Debris: Examining Altitudes, Anomalies and Attributions
