Advancing Robotics and Automation: DOE’s National Laboratory Projects

Breakthroughs in Advanced Robotics and Automation for Scientific Infrastructure

The U.S. Department of Energy’s (DOE) Office of Science has recently announced the selection of four National Laboratory-led projects aimed at driving innovations in advanced robotics and automation specifically tailored for scientific infrastructure.

Within the DOE National Laboratories, some of the world’s most cutting-edge scientific instruments, computing systems, and experimental facilities are housed and operated. These selected projects have set out to significantly accelerate advancements in advanced robotics and automation, with a focus on customizing these innovations to suit the unique research environments found within these laboratories. The ultimate goal is to create solutions that can be widely applied across various autonomous scientific operations.

DOE’s expectations from these projects include the creation of open software and interfaces, reusable robot skills, digital-twin environments, benchmark tasks, datasets and trained models, safety practices, provenance records, and training resources. The primary objective is not only to showcase four impressive robotics demonstrations but to establish capabilities that can be leveraged and expanded upon by the broader DOE scientific community. Projects were selected based on criteria such as building reusable infrastructure, advancing learning-enabled autonomy, demonstrating transferability across instruments or laboratories, integrating robotics with Super Intelligence (SI) and advanced computing, and complementing other selected projects.

Overview of Selected Projects:

1. Modular Autonomous Experimentation through Self-improving Testbeds for Robotic Operations (MAESTRO)

Argonne National Laboratory leads the MAESTRO project, which aims to develop modular, self-improving infrastructure for laboratory robotics. The approach involves a combination of digital twins and world models, reusable robot skills, agent-based orchestration, safety mechanisms, and learning from operational experience. The project will utilize a central robotics sandbox and multiple science-facing proving grounds to assess whether robotic capabilities can be enhanced through usage and transferred across instruments, workflows, and laboratory environments.

2. DOE AI Robotics Testbed to Generalize Autonomous Science (DART)

Brookhaven National Laboratory is spearheading the DART project, which focuses on determining the reliability, transferability, and scientific validity of robotic autonomy. The project will work on developing adversarial digital counterparts, a verified failure atlas, multimodal robot-learning methods, and strategies for transferring task knowledge across different robotic platforms. DART’s evaluation criteria go beyond mere completion of robotic motions to ensuring that the resulting physical and scientific outcomes meet the experiment’s requirements.

3. Testbed for Robotics and Autonomy in Connected Experiments (TRACE)

Oak Ridge National Laboratory is leading the TRACE project, building on Oak Ridge’s in-house INTERSECT automated laboratories ecosystem. TRACE aims to interconnect robots, scientific instruments, SI agents, digital twins, data systems, and computing resources through standardized interfaces. The project will focus on developing typed capability contracts, reusable sensorimotor skills, end-to-end provenance, safety and mixed-fidelity testing, and independently rerunnable evaluation methods. The emphasis lies in enabling autonomy solutions to be tested, replicated, and implemented across different laboratories and scientific domains without the need to rebuild the entire software stack.

4. A Source-to-Discovery Platform for Instrumentation, Robotics, and Embodied AI in DOE Photon-Science Facilities (SPIRE)

SLAC National Accelerator Laboratory leads the SPIRE project, which aims to integrate physical sample manipulation, intelligent detectors, accelerator and instrument controls, persistent experiment state, and edge-to-high-performance computing. SPIRE is designed to demonstrate source-to-discovery autonomy within photon- and electron-science facilities, where decisions must span timescales ranging from microseconds to hours. The project will also develop methods, interfaces, digital twins, and benchmark tasks for potential adaptation in other DOE scientific User Facilities.

Collectively, these four projects cover different yet interconnected stages of the autonomous-science lifecycle:

• MAESTRO: Developing capabilities that learn and improve through experience;
• DART: Stress-testing these capabilities and assessing their reliable transferability;
• TRACE: Connecting, packaging, and replicating capabilities across laboratories; and
• SPIRE: Deploying and validating autonomy in significant scientific-facility operations.

These projects are funded through the Advanced Scientific Computing Research program under the solicitation on Robotics and Automation Testbeds for Autonomous Scientific Discovery (LAB 26-3601), with a total funding of $30 million. $2 million is allocated for Fiscal Year 2026, with outyear funding subject to congressional appropriations. Further information on the selections can be found on the Office of Science funding page.

It’s important to note that selection for award negotiations does not guarantee funding from DOE. A negotiation process will take place between DOE and applicants, with the possibility of DOE canceling negotiations and rescinding the selection for any reason during this phase.