The Remora Mission: How Starfish Space Proved Autonomous Satellite Docking With a Single Camera

Rendezvous and proximity operations — the art of maneuvering one spacecraft close to another — is one of the most difficult things to do in space. It requires precise orbital mechanics, real-time navigation, and the ability to match the position and velocity of a target that may be tumbling, uncooperative, and difficult to see. Historically, only three entities had demonstrated RPO capability: NASA (Apollo, Space Shuttle, ISS), ESA, and JAXA. Each mission required years of planning, teams of ground controllers, and budgets measured in hundreds of millions of dollars.

In December 2025, Starfish Space and Impulse Space completed the Remora mission — and made RPO look almost routine. A small spacecraft autonomously navigated to within meters of another object in low Earth orbit using a single lightweight camera, closed-loop guidance software, and no human intervention during the final approach. The mission was the first fully autonomous commercial rendezvous and proximity operations demonstration in history, and it validated the technical foundation for every satellite servicing contract Starfish has in its portfolio.

What Made Remora Different

Previous RPO missions — from NASA's Gemini program in the 1960s to Northrop Grumman's Mission Extension Vehicle in 2020 — relied on extensive ground control. Human operators monitored telemetry, commanded maneuvers, and made real-time decisions as the servicer approached its target. This approach works for individual high-value missions but cannot scale to a market where thousands of satellites need servicing. If every approach requires a team of ground controllers working around the clock, the economics of commercial satellite servicing break down.

Remora demonstrated a fundamentally different approach. The spacecraft's onboard CETACEAN navigation system — developed by Starfish — used computer vision algorithms to identify and track the target using a single camera. Navigation filtering techniques processed the visual data to compute the spacecraft's relative position and velocity in real time. Closed-loop guidance software used this information to compute and execute maneuvers autonomously, adjusting the approach trajectory as conditions changed. The ground team monitored but did not control the approach.

The CETACEAN Navigation System

CETACEAN is the technical core of Starfish's competitive advantage. The system combines three capabilities: visual target acquisition (identifying and tracking an object using camera imagery), relative navigation (computing the servicer's position and velocity relative to the target with sub-meter accuracy), and trajectory planning (computing and executing approach maneuvers in real time). Each capability has been demonstrated individually in laboratory settings, but Remora was the first time the full system operated autonomously in orbit.

The use of a single camera — rather than the multi-sensor suites (lidar, radar, multiple cameras) used by competitors — is a deliberate design choice that reduces mass, power consumption, cost, and failure modes. A camera is passive (it does not emit signals that could interfere with other spacecraft), lightweight (grams rather than kilograms), and low-power (watts rather than tens of watts). The computational challenge of extracting precise 3D navigation data from 2D camera images is significant, but Starfish's algorithms have demonstrated sufficient accuracy for the docking and servicing missions in their contract portfolio.

The Impulse Space Partnership

Remora was a collaborative mission with Impulse Space, the orbital transportation company founded by former SpaceX Propulsion CTO Tom Mueller. Impulse provided the maneuverable spacecraft platform that served as the testbed for Starfish's navigation and guidance systems. The partnership was strategic: Starfish needed an orbital vehicle capable of the precise maneuvers required for RPO testing, and Impulse needed to demonstrate its spacecraft's agility and precision in a high-profile mission context.

The collaboration also previews a future where orbital servicing involves multiple specialized vehicles working together. An Impulse Space tug might deliver an Otter servicing vehicle to the orbital neighborhood of its target, and the Otter would handle the final approach and docking autonomously. This division of labor — long-range transportation by one vehicle, precision proximity operations by another — mirrors terrestrial logistics where different vehicles handle different segments of a delivery route.

From Demonstration to Operational Missions

Remora's significance extends beyond the technical demonstration. It de-risked the autonomy software that Starfish will use on every operational Otter mission. The three Otter vehicles scheduled for launch in 2026 — serving NASA, the U.S. Space Force, and Intelsat — will all rely on CETACEAN for their approach and docking phases. The Remora data provides flight heritage for the navigation algorithms, reducing risk for customers who are entrusting expensive satellites to Starfish's autonomous systems.

For the U.S. Space Force's PWSA disposal missions, autonomous RPO is not just convenient — it is operationally necessary. The PWSA constellation will include hundreds of satellites across multiple orbital planes. Disposing of each satellite at end of life using ground-controlled approaches would require enormous ground infrastructure and personnel. Autonomous disposal vehicles that can independently navigate to, dock with, and deorbit target satellites transform the economics from unworkable to viable.

What Comes Next

The Remora mission validated rendezvous and proximity operations with a cooperative target — a spacecraft designed to be approached. The harder challenge, and the one Starfish must eventually solve for debris removal, is RPO with uncooperative targets: dead satellites that may be tumbling, have protruding appendages, or lack any features designed for docking. Starfish's approach — using visual navigation that does not require cooperative markers on the target — is inherently better suited to uncooperative targets than systems requiring reflectors, beacons, or docking plates.

The roadmap from cooperative to uncooperative RPO will likely proceed through the Otter missions in 2026–2027, each adding complexity and capability. By the time Starfish undertakes its first uncooperative debris removal, the CETACEAN system will have accumulated multiple missions of flight heritage — a critical trust-building process in an industry where failure means creating more debris rather than removing it.

Frequently Asked Questions