How to Find PhD Supervisors in Robotics (and Get Funded)

Robotics looks, at first glance, like it should carry the same export-control story we've already covered for aerospace engineering, given how much DARPA and defense money flows into autonomy and perception research. It doesn't, not as the default. DARPA's own published policy for academic researchers says it deliberately keeps fundamental research awards free of publication restrictions, export controls, and citizenship requirements, and the core NSF program that funds robotics research, the National Robotics Initiative, is a standard research grant to a professor rather than a citizenship-gated personal fellowship. What actually makes a robotics PhD search different isn't a funding restriction, it's that the field barely has its own department: you're searching across computer science, mechanical engineering, and electrical engineering faculty listings simultaneously, often at the same school, for professors doing recognizably the same kind of work under three different administrative roofs.

Robotics isn't one department, it's three

Most schools don't have a robotics department you can simply apply to. Georgia Tech's robotics PhD requires applying through a home academic unit, options include Interactive Computing and Mechanical Engineering, with the final admission decision made jointly by that unit and the school's Robotics PhD Program Committee. The University of Utah houses its robotics PhD under mechanical engineering. Other programs house theirs under electrical and computer engineering instead. The University of Michigan's robotics program recommends, but doesn't strictly require, an engineering background, and notes that students without one are at a disadvantage early on, while guidance from smaller robotics programs describes competitive applicants as having a background in at least two of the three core areas: computer science, mechanical engineering, and electrical engineering. The practical consequence is that a robotics professor search has to cover three faculty directories at a given school, not one, and which department you're nominally housed in can determine your assistantship norms and which fellowship subfield code applies to you.

The field's core federal funding is explicitly designed to stay open

This is the real departure from the aerospace and nuclear posts in this series. DARPA's guidance for academic researchers states plainly that it keeps its fundamental research awards free of publication restrictions, export controls, and citizenship requirements, applying restrictions only when necessary for national security, or in rarer cases involving military systems or manufacturing technologies considered critical to defense, and that each solicitation states outright whether a given activity counts as fundamental research. The NSF's National Robotics Initiative, now in its third iteration as NRI 3.0 (NSF 21-559), is a multi-agency program run jointly with USDA, NASA, DOT, NIH, and NIOSH that funds fundamental research advancing robot integration, structured as a standard research grant to a professor's institution rather than a named, individually-applied-for fellowship. Neither of these facts means every robotics project is open to international students by default, but the stated default posture of the field's two largest funding sources is to stay inside the fundamental research exclusion, which is a different starting point than aerospace, where NASA's own graduate fellowship is closed to international students outright and a meaningful share of the field (satellite work specifically) has been pulled under tighter export control by design.

The exception still exists, it's just a narrower slice of the field

"Mostly open by default" isn't the same as "always open." The same mechanism covered in our aerospace engineering field hub still applies here: a sponsor's contract can strip a project of its fundamental research protection by restricting foreign-national participation or requiring pre-publication review, and disclosing controlled technical data or source code to a foreign national, even inside a US lab, can itself count as a regulated "deemed export." For robotics specifically, this risk concentrates in certain autonomy, perception, and hardware projects tied to defense sponsors with those kinds of contract terms, not across the field generally. Don't ask a professor whether their lab does defense-funded work, most labs that do plenty of openly published research also hold some defense funding. Ask which specific project you'd be joining, and whether that project's funding agreement includes any foreign-national or publication-review restriction.

GRFP covers robotics as a subfield, but the citizenship rule still applies

University-maintained lists of NSF Graduate Research Fellowship Program eligible majors list "Robotics and Computer Vision" as a named subfield under Computer and Information Science and Engineering, so the field itself is recognized. That doesn't change GRFP's underlying eligibility rule: applicants must be US citizens, nationals, or permanent residents by the application deadline, with no exception for field or academic record. For an international applicant, GRFP is closed the same way NASA's NSTGRO is closed in aerospace, the difference in robotics is simply that the professor-held funding you'd rely on instead, DARPA fundamental research awards or an NRI-funded research assistantship, carries less of a built-in citizenship restriction than the equivalent professor-held funding in a field like nuclear engineering, where even non-fellowship national-lab access can introduce its own separate citizenship question.

Industry labs are a real, verifiable third funding channel

Robotics also has a funding channel that's far less prominent in the aerospace and nuclear posts in this series: direct, large-scale industry sponsorship of academic labs, structured around open publication rather than product secrecy. Toyota Research Institute's university research program is the clearest verifiable example: in a 2021 round, it funded $75 million over five years across 16 universities for AI research spanning automated robotics and machine-assisted cognition, with Penn's GRASP Lab among the recipients. By 2023, TRI stated its academic program had funded $100 million in research and generated more than 1,250 paper submissions since it began in 2016. That's a meaningfully different funding relationship than a defense contract: it's sponsored research built around publishing results, not around restricting who can see them, though you should still confirm a specific lab's actual terms directly rather than assuming every industry-funded robotics project works the same way.

A practical search order

• Search faculty listings in computer science, mechanical engineering, and electrical/computer engineering at a given school, not just one department, before concluding a school "doesn't do" the robotics subfield you're interested in.
• Check whether a specific professor's funding comes from DARPA, the NSF's National Robotics Initiative or another core NSF program, industry sponsorship (Toyota Research Institute's university program is a verifiable example), or a named, citizenship-restricted fellowship like NSF GRFP, each implies a different eligibility picture for you as an applicant.
• If the lab does defense-sponsored autonomy, perception, or hardware work, ask specifically which project you'd join and whether its funding agreement includes a foreign-national or publication-review restriction, rather than assuming the lab's defense funding in general makes it closed, or that its open-publication reputation makes every project inside it open.
• If you're a US citizen or permanent resident, GRFP's "Robotics and Computer Vision" subfield and other named engineering and computer science fellowships remain genuinely worth applying to, and can strengthen your case with a professor who already holds related grant funding.

GradScoutFunding handles the professor-identification step across however many departments a school splits robotics across: it searches professors by field and country, confirms they're actively publishing, and drafts a personalized first email grounded in their real recent paper for you to review and send yourself, it never sends anything automatically. The free tier gives 100 credits with no card required, and paid credit packs are one-time purchases that never expire.

Common questions