Worcester Polytechnic Institute Engineers Develop Bat-Inspired Drones That ‘Hear’ Through Fog and Smoke Using Ultrasound

Worcester Polytechnic Institute researchers develop palm-sized drones using bat-inspired ultrasound to navigate through fog, smoke, and total darkness.

By: AXL Media

Published: Mar 28, 2026, 1:16 PM EDT

Source: Information for this report was sourced from Worcester Polytechnic Institute

Mimicking Nature’s Echolocation

Researchers at Worcester Polytechnic Institute have successfully developed a six-inch quadrotor capable of navigating complex environments without the use of cameras. Led by Assistant Professor Nitin J. Sanket, the team drew inspiration from the biological efficiency of bats, which navigate cluttered, dark spaces using minimal neural processing. While traditional drones rely on heavy Lidar or power-hungry vision systems, the WPI drone uses two tiny ultrasound sensors and deep learning to interpret echo patterns. This "acoustic vision" allows the robot to identify and avoid obstacles by "hearing" the environment, significantly reducing the computational load required for autonomous flight.

Breaking the Visual Barrier

Traditional drone navigation is often rendered useless in search-and-rescue scenarios involving fire, fog, or heavy snow. Visual sensors struggle with low contrast and occlusions, while Lidar can be heavy and expensive. The WPI team’s X-shaped drone, weighing approximately one pound, was specifically tested against these limitations. During 180 trials, the drone demonstrated a success rate of up to 100% while navigating through simulated hazards, including thick fog and complete darkness. This capability is critical for first responders who need to deploy small, agile robots into collapsed buildings or smoke-filled corridors where human visibility is zero.

Overcoming Self-Generated Noise

One of the primary hurdles in acoustic drone navigation is the noise generated by the drone's own propellers, which can drown out the subtle echoes needed for echolocation. To solve this, the WPI researchers engineered a specialized acoustic shield that dampens internal noise, allowing the sensors to detect incoming ultrasound waves clearly. By combining this hardware solution with a deep-learning onboard computer, the drone can distinguish between its own mechanical sounds and the reflections from nearby walls or obstacles. This breakthrough enables the drone to operate effectively in "noisy" environments that would typically baffle sound-based sensors.