Traditional drone navigation systems rely on pre -mapped or slow real -time calculations, limiting their ability to react dynamically. However, researchers from the University of Hong Kong have developed Fast aviation work (great) that is on safety (great) -MAV compact with a 280-mm wheelbase and rush ratio to more than 5.0. This autonomous new generation drone can make decisions in a split second with a perception based on Lidar, real -time mapping and a two -level strategy.
Its advanced navigation system allows great to move with complex environments at high speeds, while avoiding real -time obstacles. Integrating the detection of Lidar and the intelligent planning framework, the drone provides both agility and safety – even in a completely unknown area.
The key innovation standing for great is its two -level planning system, which allows him to discover new paths, while maintaining a spare plan for safety:
- Exploration trajectory – presents a fast, efficient path to the target, even through unknown spaces.
- Backup trajectory – the drone provides always to return to a known safe space if the exploration path encounters obstacles.
This system again calculates the trajectory 10 times per second, enabling great response to changes in the environment and avoiding collisions – even at high speeds. If the replan fails, it will perform the last involved trajectory – guaranteeing safety, ensuring that the drone will never enter unknown or dangerous areas without a way back.
One of the biggest challenges in real -time drone navigation is quickly and accurately understanding the environment. This is a great system with a sophisticated module of perception based on Lidar:
- The Lidar Drone sensor gives 200,000 points per second, detecting obstacles at a distance of 70 meters.
- The dynamic mapping system ensures storage of only relevant data, preventing the mess that could slow down the calculations.
- The strategy of the sliding cloud removes outdated information (e.g. moving objects that are no longer present), preventing false detection of obstacles.
This level of perception allows you to super effectively navigate through forests, urban environments and disaster zones, in which obstacles are unpredictable and are constantly changing.
What makes this navigation system really unique is to generate a spare corridor in real time-a power to safety that the drone provides always has a collision-free escape route.
Using the innovative CIRI algorithm (iterative regional inflation, iterative configuration), the system separates secure flight corridors directly from Lidar data. Unlike older methods, which are based on pre -mapped environments or expensive computation, CIRI allows the production of safe zones in real time, significantly reducing the general costs of processing.
- Ciri ensures that each spare corridor is fully contained in a known free space, preventing navigation failure.
- The system accumulates the last Lidar scans (1-2 seconds) to improve the corridor accuracy.
- Even with a limited field of view, Ciri dynamically adjusts the backup zone to guarantee safe redirect.
By eliminating the need to map radiation, this breakthrough enables quick UAV operations in difficult and unpredictable environments.
The ability to generate safe trajectories to create backups in real time means a serious forward leap in drone autonomous technology. This innovation is ideal for:
- Autonomous deliveries of drones in complex urban environments.
- Searching for mission in disaster zones.
- Military recognition at unknown or high risk.
- Precise agriculture in which UAV scans large fields with a minimal delay.
Combining a fast flight, real -time mapping and intelligent navigation, it super exceeds the boundaries of what autonomous drones can achieve. As these possibilities evolve, we can soon see drones navigating complex environments with the same agility and awareness as real pilots – if not better.
