UC Drone System Uses Audio to Detect Motor Failures Before Takeoff

65% of North American warehouses still rely on manual inventory processes, a gap that University of Cincinnati doctoral students are addressing with autonomous drone systems. In a separate project, the same engineering team is developing audio-based diagnostics for NASA to identify motor and propeller damage before takeoff.

The research began in 2024 when NASA expressed interest in detecting mechanical failures prior to flight. Prissha Chawla, a doctoral student in the College of Engineering and Applied Science, led the effort to record the specific sounds emitted by drones with chipped or cracked propellers and failing motors. The team converted these acoustic profiles into data that ground station computers can analyze. This allows operators to receive an alert if a drone is deemed not airworthy, preventing potential crashes in public safety operations.

Chawla noted that grit and debris often enter motor bearings, causing them to seize. The system focuses on the pre-flight tests that NASA already conducts. The long-term goal is to create an autonomous drone that can listen for its own mechanical failures in flight and return to base before a catastrophic breakdown occurs.

Simultaneously, doctoral students Deepak Antony David and Luke Busse are working with Valvoline on a warehouse automation system. Their drones use light detection and ranging, or LIDAR, to navigate indoor spaces with centimeter-level accuracy. This technology is more common in geology but is being adapted for tight warehouse aisles. The drones do not rely on satellite signals, making them suitable for indoor environments where GPS is unavailable.

The system utilizes a digital twin, which is a photorealistic 3D model of the warehouse floor. Workers can view this virtual space in high-definition video. By clicking on a specific aisle or section, they dispatch a physical drone to that location to perform tasks such as locating items or inspecting damaged pallets. Busse described the interface as having a high fidelity that resembles a video game. The virtual reality system runs on Unreal Engine software, the same platform used in titles like Fortnite and Batman: Arkham City.

The students presented both projects at the Modelling, Estimation and Control Conference in Phoenix, Ariz., this month. The conference was sponsored in part by UC's College of Engineering and Applied Science. National media outlets, including Yahoo! News and Tech Xplore, covered the student work following the presentation.

The warehouse project addresses a critical inefficiency in supply chains. Busse stated that manual inventory picking leads to categorization errors and misplaced products, which take time to correct. Drones can reach high shelves quickly and navigate tight confines without difficulty. The team refined the system in a high-bay lab at Digital Futures, where they set up a mock warehouse with pallets of empty boxes. The drone captured the space's contours and sent a real-time 3D rendering to a laptop.

Demand for such automated systems is rising as retailers build new facilities. Analysts at Interact Analysis estimate that more than 4,000 new warehouses will be constructed in the United States over the next four years. The UC team works through the Industry 4.0/5.0 Institute, which partners with companies to solve industrial problems using emerging technology. The audio damage detection system could eventually become a standard feature in drone-in-a-box deployments, where sensors in the docking station provide preflight checks.