A Worcester Polytechnic Institute researcher is building palm-sized robots that mimic bat echolocation to navigate harsh search and rescue environments. These AI-powered drones use ultrasound sensors and custom 3D-printed sound modulators to detect obstacles within two meters, potentially replacing human searchers in dangerous conditions like smoke, dust, and extreme weather.
Search and rescue missions just got a potential game-changer that fits in your palm. Professor Nitin Sanket at Worcester Polytechnic Institute is developing bat-inspired robots that could replace human searchers in the most dangerous environments.
The palm-sized flying robots use ultrasound navigation - just like bats - to spot obstacles within a two-meter radius. What makes them special isn't just their size, but how they solve the fundamental problem that's plagued miniature drones: noise interference from their own propellers.
"Search and rescue is done on foot," Sanket told TechCrunch. "There are a lot of people who go on foot with flashlights in really harsh conditions and put their lives at danger to save others. We thought drones are the answer because they can cover a lot of ground really fast."
The breakthrough came from studying how bats handle their own flight noise. These nocturnal mammals have specialized tissues in their nose, ears, and mouth that adaptively change thickness and density to modulate sound. Sanket's team replicated this with custom 3D-printed structures placed in front of each robot.
"Bats have these special tissues which essentially change the shape of the sound itself," Sanket explained. The artificial version does the same job - filtering out propeller noise so the ultrasound sensors can actually detect obstacles and navigate safely.
This isn't Sanket's first dive into biomimetic robotics. His fascination with nature-inspired engineering started during his PhD, when his advisor challenged him to build the smallest robot possible. That led to an ambitious prototype: a robotic beehive of tiny drones designed to pollinate flowers. While that project proved too ambitious for immediate deployment, it sparked the research methodology that's now yielding practical results.
"We had to reimagine what a drone would be at that point, which is go back to biology, because biology does this way better than we can today," Sanket said. "How do insects or birds do it with super limited compute and not-so-good sensing apparatus?"
The current robots use ultrasound sensors borrowed from automatic faucets - a clever choice that keeps power consumption minimal while providing reliable obstacle detection. AI-powered software filters the incoming ultrasound data, separating useful navigation signals from background noise.
