A robotic system commissioned by DARPA has, for the first time, built a truss structure directly in space. This opens the way for using artificial intelligence to construct extraterrestrial stations and colonies. However, that remains a distant prospect.
A Robotic Builder in Space
A robotic truss builder developed, designed and assembled by researchers at the California Institute of Technology (Caltech) has successfully constructed a structure in space—for the first time without help from astronauts. Part of DARPA’s Novel Orbital and Moon Manufacturing, Materials, and Mass-efficient Design (NOM4D) programme, the truss builder consists of an autonomous manipulator that retrieves struts and joints from nearby dispensers and repeatedly assembles identical sections, or bays, of a perimeter truss. The website phys.org reports on the project.
“This is the first time a structure has been built autonomously in space,” says Sergio Pellegrino, professor of aerospace and civil engineering, holder of the Lynn Booth and Kent Kresa Leadership Chair in Aerospace, executive officer for aerospace, and senior research scientist at JPL, which Caltech manages for NASA.
Pellegrino leads the team behind Caltech’s Autonomous In-Space Tensioned Reflector Assembly (ASTRA) demonstration project. The robotic builder has spent the past six months in space, working to assemble a reflector antenna measuring 1.4 metres (4.6 feet) in diameter from fibre-composite tubes, aluminium joints and a Kapton reflective surface.
Although the team faced time constraints and could build only the initial portion of the antenna, the demonstration established the feasibility of the proposed assembly concept, which could be scaled up to very large perimeter trusses supporting giant reflective surfaces.
A Demonstration in Space
The space demonstration is the third phase of the DARPA project. During the first two phases, the ASTRA team developed the concept and design methodology, then demonstrated that the concept worked under laboratory conditions on Earth. The space payload was completed, tested and delivered to its hosting provider, Momentus Inc., in less than 12 months.
The scientists say that developing such complex hardware in such a short time was an enormous undertaking. Now, with the demonstration completed, they can begin considering new research directions.
Scientists would very much like to be able to deploy large structures, such as reflective surfaces, in space. These could serve as enormous mirrors for powerful space telescopes, allowing astronomers to observe even more distant objects, or form the apertures of large radio-frequency antennas.
Autonomous Assembly of Very Large Structures in Space
The size of such structures is currently limited by the dimensions of rocket fairings. Since the beginning of space exploration, there have been two ways to deliver large structures into orbit or beyond: fold them for launch and deploy them in space, or assemble them manually from components delivered over a series of launches. A third option is now needed to place larger structures in space: build them there.
This is where the NOM4D programme comes in. DARPA launched the programme in 2022 with the ultimate goal of “delivering raw materials from Earth for manufacturing in orbit,” according to the programme’s website. Caltech’s ASTRA demonstration is one of only two projects to reach the third phase of NOM4D: small-scale demonstrations in orbit.
The other group, led by the University of Illinois Urbana-Champaign, will use a module attached to the International Space Station to test a method for strengthening carbon-fibre tubes in space.
“The goal of NOM4D is to enable efficient, scalable construction of very large structures in space,” Andrew Detor, DARPA’s NOM4D programme manager, said in a written statement. “By ‘very large,’ I mean hundreds of metres or kilometres in scale. Nothing approaching that scale is achievable today.”
ASTRA in Orbit
Early on March 30, 2026, ASTRA departed for low Earth orbit from Vandenberg Space Force Base as one of 10 payloads carried by Momentus Inc.’s Vigoride 7 orbital spacecraft, which itself launched on SpaceX’s Transporter-16 mission. The ASTRA payload, occupying approximately one cubic metre (35 cubic feet), survived the launch.
It contained two preassembled truss bays with diagonal braces made from structural struts and joints, along with an arm-like manipulator. The payload also carried all the components needed to build the rest of the truss, stored in motorised dispensers.
Once ASTRA was in space, the team began sending commands to its onboard computer through the Vigoride spacecraft. First, the launch restraints holding its components in place during launch were released. Next, the two preassembled bays were pushed out of the truss builder, followed by the assembly of a third bay.
Commands to build the structure were sent by ASTRA team members Charles Sommer, a former systems engineer at Caltech; Alan Truong, a Caltech research scientist; and graduate student Tara Venkatadri (MS). The results were checked using photographs downloaded from space.
The operation was repeated to overcome unforeseen problems, of which there were many. “That is exactly why you conduct proof-of-concept testing. Conditions in space differ from those on Earth in so many ways,” explains Jayden Suh, a former postdoctoral researcher at Caltech. “There are always lots of little surprises.”
The Truss Assembly Process in Space
To help resolve those little surprises, the team built an ASTRA twin—a flight spare they could work with in the Caltech laboratory. There are only about seven or eight brief windows of 15 minutes each day during which researchers on the ground can communicate with the system in space. The team therefore finds solutions by testing alternatives on the spare, then sends the workaround to its twin in orbit.
In normal operation, the main robotic manipulator moves around, picks up components, positions them and assembles each bay on an assembly plate. Once this sequence is complete, the manipulator attaches a mesh-like perforated film to the truss to form two reflective cable nets, which are preloaded against one another using springs.
ASTRA’s reflective material was jointly invented by the Caltech team and collaborators at the University of Pennsylvania led by Professor Igor Bargatin. The perforated film has several advantages over the conventional knitted metal mesh typically used in deployable reflectors: it requires less preload, allowing engineers to design much lighter structures.
Once fully prepared, each new bay is pushed out into space, and the entire sequence is repeated for the next bay. ASTRA’s truss is considered complete once 11 bays have been built, attached to the mesh reflectors and pushed outward.
With ASTRA’s mission ending in December, the demonstration is unlikely to complete the entire truss assembly. However, the team says it has gained enough experience to show that the proposed approach to in-orbit assembly is feasible, while also identifying the need to simplify and refine the design.
“When you reach the size limit beyond which structures cannot be preassembled on the ground and deployed, you have to think about how to build them in space,” Pellegrino says. “So this work could be useful not only for astronomy and astrophysics, but also for communications and other applications, such as building very large solar arrays for space-based solar power.”
