CMU Collaborates With Local Communities To Design Safer Robotic Systems NSF Funds Development of Robots for Public Spaces
Amanda SapioThursday, October 2, 2025Print this page.
The Breakdown
- We need research into how humans and robots can interact safely in public.
- CMU researchers will involve community members in designing these robotics systems.
- The NSF is providing $1.25 million to support the work.
***
A Carnegie Mellon University research team recently received a $1.25 million grant from the National Science Foundation's Smart and Connected Communities program to design robots that can operate in areas such as sidewalks, street crossings and parks without risking human safety.
As robots increasingly integrate into day-to-day human activities, questions abound about how the rules surrounding public spaces could change. And, perhaps more importantly, how new technologies will affect community safety overall.
Led by Nikolas Martelaro, an assistant professor at the Human-Computer Interaction Institute (HCII) in CMU's School of Computer Science, the CMU research team brings together researchers from across the university, including Sarah Fox, also an HCII assistant professor; Ding Zhao, an associate professor in the Mechanical Engineering Department; and Corey Harper, an assistant professor in the Civil and Environmental Engineering Department. This multidisciplinary approach allows each team member to address the research from a unique perspective.
The team will also collaborate with the Lawrenceville Corporation, a local community organization in Pittsburgh; the Urban Robotics Foundation; and Carnegie Mellon's Metro21: Smart Cities Institute. The researchers plan to work with community members to brainstorm scenarios where robots could encounter people on public walkways, design navigation and communication systems, and test the resulting strategies in the group's lab on CMU's campus. The team wants to give community members an opportunity to describe sidewalk scenarios they might be concerned about and how they expect robots to act in those situations, including how they should move, what rules they should follow and how they should talk to pedestrians.
"Community input isn't just helpful, it's essential. These robots will move through people's neighborhoods and daily routines, so residents need a genuine voice in shaping how they behave," Fox said. "It's not enough for us to design machines that operate safely. They must also reflect the values and priorities of the communities they serve. That's why community voices aren't an afterthought — they're the foundation of how we define success in this work."
Zhao said his Safe AI Lab focuses on how to safely deploy artificial intelligence to the real world, including on sidewalks.
"I'm interested in the sidewalk because the rules around sidewalk traffic are more complex than car lanes, yet more standardized than the private household environment," Zhao said. "The team will leverage existing world model technologies that can generate realistic interactive scenarios that enrich safety-critical datasets with counterfactual cases."
Although robots have great potential to enhance communities through food and medicine delivery, sanitation, safety patrol, and various other benefits, they still face significant challenges when navigating real-world sidewalks. The team will focus on teaching robots how to maneuver in different weather, traverse broken walkways and safely avoid pedestrians who may be distracted by their phones, walking pets or pushing strollers.
"I think one of the things we want to do is make sure that the community has an important role — we want them to know that they are a critical part of this research," Harper said. "We don't want to simply introduce a robot without getting their input first. We also hope that this research can be used by future communities, cities and planning organizations, especially as we see robots deployed at a larger scale. The community feedback is a very important part in ensuring the robots we deliver are in line with what the community would like to see for the future."
The team plans to train robot navigation and communication systems using 3D simulation environments, laboratory studies and field tests. The project will examine the needs and requirements of both wheeled robots, such as the delivery robots found in some U.S. cities, and quadruped robots, like a robot dog, in navigating shared spaces.
The spark for the project happened when companies first deployed sidewalk delivery robots in Pittsburgh. The researchers learned that Pennsylvania legislation considers robots to have the same walking rights as pedestrians when traversing sidewalks. They also realized these same robots can weigh up to 500 pounds and operate at 12 miles per hour.
"When we learned this, we found that to be quite heavy and fast," Martelaro said. "We also learned that robots have a lot of challenges navigating sidewalks. Sometimes they get stuck, sometimes they have awkward interactions with people. And then more and more research came out showing that there were potential safety risks. This got us thinking about the future design of these robots, both in the way they navigate and how they communicate."
The team's goal is to design robots that can blend in and navigate safely alongside people. They don't foresee a world in which robots will have dedicated sidewalk lanes or separate areas to traverse.
"We want to encourage people in the community to see our robot prototypes both in the simulation environment in our lab and in real-world environments so they can give direct feedback," Martelaro said. "This process will help us make changes based on what's going to work for the local community."
Community members have already asked great questions during some informal conversations with the researchers about this topic, ranging from what these robots will do and the rules they'll have to follow to if robots will listen to humans or bother them on the sidewalk. The researchers also heard concerns about sidewalk robots during similar past projects and at a community meeting with city officials about sidewalk robots.
"These questions have helped us think about how we can engage the community in this project and have conversations that lead to stronger human-robot interactions," Martelaro said.
In addition to the questions from the community, the team was also surprised by the individuals who have shown interest in the robotics project so far.
"I was excited to learn that many older adults are interested in this study," Zhao said. "Especially because robots are used to deploy food and medicine that otherwise would be more challenging to access."
Aaron Aupperlee | 412-268-9068 | aaupperlee@cmu.edu