US researchers create palm-sized drones using bat-inspired ultrasound and AI to navigate smoke, fog, and obstacles.
US researchers have developed palm-sized drones that use bat-inspired ultrasound and AI to navigate through fog, smoke, and other challenging conditions. The breakthrough by a team at Worcester Polytechnic Institute mimics how bats rely on simple echolocation to move through dark, cluttered environments with minimal neural processing .
Using just two tiny sensors and lightweight computation, these drones can independently sense their surroundings and make decisions, offering a low-cost, efficient alternative to heavier, camera-based navigation systems. According to the team, the innovation could enhance search-and-rescue operations in hazardous environments where traditional drones struggle.Bat-inspired navigationThe WPI team focuses on nature-inspired robotics, drawing lessons from creatures like bees and bats to design more efficient autonomous aerial robots. Typically, drones navigate using a combination of sensors, cameras, controllers, a power source, and sophisticated algorithms that interpret environmental data to make decisions. Many rely on technologies such as lidar or radar, which analyze light or radio waves to map surroundings. While effective, these systems are often heavy, costly, and consume significant power, limiting the drone’s flight time and maneuverability. Additionally, visual-based systems struggle in darkness, poor weather, or smoky conditions, and even slight noise from drone propellers can interfere with sound-based perception. Processing this information consumes extra energy, further reducing efficiency.To address these challenges, Sanket’s team developed a compact, X-shaped quadrotor drone roughly six inches wide, equipped with lightweight ultrasound sensors. The drone also incorporates an acoustic shield to dampen noise from its own propellers, enabling clearer detection of echoes. Using deep learning, the team trained the drone’s onboard computer to interpret subtle ultrasound echo patterns in a manner inspired by bat echolocation, allowing the robot to “hear” obstacles and navigate effectively with minimal computation.The researchers tested the drone both outdoors, in a wooded environment, and indoors, in a laboratory filled with challenging obstacles, including transparent plastic and metal poles. Despite weighing only about one pound, the drone successfully maneuvered through cluttered spaces, demonstrating the potential of ultrasound-guided flight for small, low-power aerial robots operating in environments where traditional navigation systems often fail.Obstacle-sensing flightThe team conducted indoor tests under challenging conditions, including complete darkness with black obstacles and simulated environmental hazards such as fog and snow blown onto the course. Despite these difficulties, the palm-sized drone was able to navigate autonomously, with enough battery power for about five minutes per flight. Across 180 trials, the drone achieved a reported success rate ranging from 72 percent to 100 percent, demonstrating its ability to maneuver through cluttered and visually obscured environments.However, the system faced limitations when encountering very thin or slender objects. Metal poles and narrow tree branches, which reflected weaker ultrasound signals, were more difficult for the drone to detect and avoid, highlighting areas for future improvement.Looking ahead, the team envisions smaller, lighter versions of the drone that could take full advantage of the low-power, ultrasound-based system, allowing for longer flight times. Enhancements could also increase maneuverability and flight speed, further expanding the practical applications of bat-inspired drones. This research points toward a future in which compact, energy-efficient aerial robots can operate effectively in conditions that challenge traditional visual or radar-based navigation systems.“In a real search-and-rescue mission, a few more seconds of flight time could mean the difference between life and death for a survivor,” said Nitin J Sanket, assistant professor in the Department of Robotics Engineering at WPI, in a statement on TechXplore.
AI Model Bat-Inspired Drones Drone Navigation Neural Processing Sensor Ultrasound Worcester Polytechnic Institute
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
$100 million? $200 million? What’s an elite Palm Beach mansion really worth?Price-wise, Palm Beach’s megamansion scene is addressed to the nines.
Read more »
The Blue Tsunami Hit Mar-a-Lago. Surely Trump Is Sunk, Right?A Democrat’s victory in Palm Beach County, Florida, must rankle the president, which is all the excuse I need to write about it.
Read more »
Busy DC schedule, Trump says, drove his Florida mail vote even after Palm Beach tripsPresident Donald Trump said Thursday that he voted by mail in Florida elections because his duties as president required him to remain in Washington.
Read more »
This newly redeveloped West Palm Beach neighborhood is loaded with New York spiritMeet Palm Beach Island’s little rad rising ‘hood.
Read more »
These are the hottest new restaurants in Palm Beach in 2026Order up from the waiter like you’re a local alligator.
Read more »
Trump Offers Wild Excuse for Wartime Mar-a-Lago TripWhile war raged in Iran, Trump golfed in Palm Beach.
Read more »