Drones get new ultrasonic vision upgrade, thanks to bat-inspired AI

Researchers at the University of Michigan have developed an AI-powered echolocation system inspired by bats and dolphins that enables drones and robots to navigate in complete darkness without relying on cameras, GPS, or laser sensors. This ultrasonic vision technology uses high-frequency sound pulses and analyzes their echoes to create spatial maps of surroundings, allowing machines to “see” through smoke, dust, or blackouts. Funded by the US Army Research Office and Ground Vehicle Systems Center, the system is particularly suited for disaster zones and hostile environments where traditional vision-based tools fail. The AI model employs an ensemble of convolutional neural networks (CNNs), each trained to recognize specific object shapes from echo patterns, enabling modular learning without retraining the entire network. The system was trained entirely in a synthetic 3D virtual environment simulating real-world distortions, which reduced development costs and time while maintaining accuracy. Tests demonstrated the AI’s ability to distinguish between similar echo patterns from different objects, proving its robustness in complex scenarios. Beyond defense and robotics