Drones could gain 'pain' to predict instability before failure

Researchers from Delft University of Technology and Wageningen University & Research have developed a novel approach to enhance the safety of autonomous systems like drones. Inspired by natural systems that exhibit 'critical slowing down' as they approach a tipping point, the team has demonstrated that similar early warning indicators can predict instability in engineered systems.

This concept is akin to how humans experience pain after an injury, providing feedback that prompts adjustments to prevent further damage. "Machines generally lack this form of self-awareness," explained Jasper van Beers, a researcher at Delft University of Technology. "The new indicators, derived from real-time measurement data, offer a first step toward giving engineered systems a similar ability to recognize when they are approaching their limits."

The study, published in Proceedings of the National Academy of Sciences, applies indicators based on critical slowing down, a phenomenon where systems take longer to recover from disturbances as they become less resilient. While widely used in ecology, its applicability to actively controlled systems like drones was previously uncertain due to their real-time controllers. However, the researchers found these ecological warning signals reliably predict approaching instability in such systems.

Validation occurred at Delft's CyberZoo facility, where drones were intentionally pushed to their limits and damaged to collect data on failure development. A significant advantage of this method is its model-free nature, relying solely on inexpensive onboard sensors to detect subtle behavioral changes. This allows for broad application across various engineered systems, including critical infrastructure, aircraft, and autonomous vehicles, with initial impacts expected in the growing drone sector.