Researchers at Durham University have introduced CORTO-Planner, a new drone navigation system that lets autonomous aircraft fly faster, smoother and safer through crowded, obstacle-rich environments. The work targets applications like search and rescue, infrastructure inspection, forest exploration and warehouse automation.
Flexible safe corridors
Conventional planners wrap drones in rigid safety zones, forcing sharp slowdowns near obstacles. CORTO-Planner instead builds a piecewise parametric safe corridor – a flexible safety region that adapts to the surrounding geometry while remaining smooth enough for high-speed flight.
With this approach:
- flight corridors offer up to 6× more usable space than previous methods,
- clearances around obstacles are up to 59% wider,
- drones gain much more freedom to maneuver without sacrificing safety.
This directly addresses a long-standing problem in autonomous flight: moving quickly through narrow gaps, sharp turns and maze-like layouts while maintaining reliable obstacle avoidance.
Real-time planning and broader impact
The team also designed an efficient computation scheme that avoids slow, heavy optimization. As a result, CORTO-Planner runs in real time, constantly updating the route so the drone can react to changes while keeping safe distances.
Extensive simulations and real-world tests show that CORTO-Planner consistently achieves shorter flight times and higher average speeds than several state-of-the-art planners, while preserving safe operation.
Scientifically, the method bridges a key trade-off: earlier work typically chose between accurate but jerky flight corridors and smooth but overly restrictive ones. CORTO-Planner combines both advantages, yielding paths that are smooth, adaptable and suitable for high-speed autonomous navigation.
Beyond performance, smoother paths cut unnecessary motion, improving battery efficiency and reducing mechanical wear. The software has been released as open-source, and the authors note that the underlying idea could also benefit self-driving cars, robotic arms and underwater robots navigating confined spaces.
Source: Phys.org
