Select Language:
Researchers at the University of Houston have created an advanced safety system designed to help drones avoid collisions during flight. This technology continuously observes the drone’s movements and can intervene immediately if it detects the craft is on a collision course.
Developed by Marzia Cescon, an assistant professor of Mechanical and Aerospace Engineering at the University of Houston’s Cullen College of Engineering, the system was detailed in the Journal of Dynamic Systems, Measurement, and Control.
Today, drones are increasingly employed across various sectors, such as inspecting bridges, power lines, and buildings, capturing aerial photos and videos, delivering packages, and supporting emergency response efforts.
The most common drone type is the quadrotor, also called a quadcopter, which features four spinning rotors. These allow the drone to move in multiple directions, hover in place, and navigate tight spaces with high precision. Despite their capabilities, drones are still vulnerable to unexpected disruptions like sudden gusts of wind, unforeseen obstacles, or glitches in navigation software. Such issues can cause them to veer off course, potentially crashing into structures, trees, power lines, or other objects, risking damage and safety hazards for nearby individuals.
To address these concerns, Cescon developed what she refers to as a “safety supervisor.” This is a compact software component that runs concurrently with the drone during flight. Rather than replacing existing control systems, it works alongside them, continuously assessing the drone’s operational safety.
The safety system tracks key data points like the drone’s location and tilt angle, analyzing this information in real time. It predicts whether the drone is approaching danger before a collision occurs. Cescon likens the system to an invisible safety fence—if the drone remains within a safe zone, the supervisor remains passive, monitoring silently. If it forecasts a potential breach of this safety boundary, it automatically nudges the drone onto a safer flight path.
This technology relies on a mathematical framework called a Control Barrier Function, which calculates the proximity to danger and briefly takes control if necessary to prevent accidents. Although the math is intricate, its function is straightforward: to continuously assess risk and intervene when needed.
The system has been developed and tested at the University of Houston’s Advanced Learning, Artificial Intelligence, and Control Laboratory. Testing showed that it integrates smoothly with existing drone flight controllers, making it practical for real-world application.
As drones become more prevalent in transportation, infrastructure inspection, agricultural work, and emergency services, ensuring safety is paramount. A system capable of monitoring flights, reacting instantly to unforeseen issues, and averting crashes could significantly enhance the reliability of drone operations, boosting confidence among operators and users alike.
This breakthrough marks an important step toward safer autonomous drones capable of navigating complex environments with reduced risk.
ChatGPT
Perplexity
Gemini AI
Grok AI
