New system lets drones autonomously perform complex aerobatics, safely outmaneuvering humans in both indoor and outdoor settings.
A new system is giving drones the ability to perform complex aerobatic maneuvers autonomously, safely navigating both indoor and outdoor environmentsDesigned to outmatch even skilled human pilots, the system, developed by a team at Zhejiang University in China , enables drones to plan efficient, obstacle-free flight paths.
It works by breaking down desired aerial movements into solvable metrics and incorporates safety protocols throughout.Integrated with the drone’s onboard computing and visual localization tools, the framework uses preloaded maps and camera inputs to guide precise navigation during acrobatic flights.“The proposed system enables drones to autonomously achieve flight performance usually reserved for professional pilots, unlocking boundless potential for aerobatic flight evolution in uncrewed aerial vehicles,” said the researchers in a statement.Instinctive aerial intelligenceBeyond routine drone operations, aerobatic flight entails intricate, dangerous maneuvers that frequently call for unstable attitudes. However, in the wild, bats and birds use these motions intuitively to survive, demonstrating exceptional agility and flexibility in confined, challenging spaces. Drones with comparable aerobatic agility have been developed as a result of these capabilities.Although proficient first-person view pilots may accomplish these tasks by hand, fully autonomous aerobatic flight in congested areas is still difficult. Present-day aerial vehicles find it challenging to safely perform a variety of maneuvers or avoid obstructions, particularly when maneuvering through confined locations or gathering data from challenging angles. An illustration of the proposed strategy for aerobatic maneuver representation, generation, and execution.Drones can already flip and power loop thanks to earlier research, but these techniques frequently depend on preset routes or need a lot of manual adjustment.These days, sophisticated drones have powerful planning and control algorithms in addition to vision-based and LIDAR technologies for precise localization. However, trajectory optimization techniques that are unable to flexibly combine maneuvers or take attitude singularities during severe movements into account limit the application of these tools to aerobatic flight.To overcome the limits of current approaches that limit aerial performance to safe but limited movements, systems must incorporate obstacle avoidance, human-like intention, and dynamic feasibility within 3D rotating space. Only then will autonomous aerobatics reach its full potential.Aerial autonomy advancesThe new study introduced a comprehensive system enabling quadrotors to perform any aerobatic maneuver while avoiding obstacles and bypassing differential flatness singularities.The approach defines aerobatic trajectories using two elements—spatial topology and attitude variations—called “aerobatic intentions.” These are converted into solvable metrics considering safety and dynamic feasibility. A novel yaw compensation mapping ensures stability near singularities. The system was tested on two FPV drone models, using known maps and onboard autonomy for planning, localization, and control.“The framework accounts for safety measures and condenses the desired trajectories into a series of metrics that can be solved to generate maneuvers. In addition, the system integrates with the drone’s onboard computing and localization abilities to help it navigate using preloaded maps and visual inputs,” said the team in a statement.The system was tested through both simulations and real-world flight experiments, demonstrating safe and autonomous operation in confined indoor spaces and expansive outdoor settings. Drones equipped with the system successfully executed aerobatic maneuvers such as barrel rolls, power loops, and complex combinations while navigating obstacle-filled circuits.Researchers claim the results showed that the system enabled drones to complete challenging flights more safely and efficiently than those controlled by skilled human pilots.The details of the team’s research were published in the journal Science Robotics.
AI AI Framework Autonomous China Drone Drone Navigation FPV Pilot Navigation Uncrewed Aerial Vehicles Zhejiang University
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
MassDOT eyes a new frontier for drones: medical deliveriesMassDOT says drones could be a faster, cheaper alternative to ground deliveries. State officials have had preliminary talks with health care organizations about using them, following a series of successful test flights last fall..
Read more »
Russia leveling up Shahed drones with new Chinese antennae: UkrainiansBusiness Insider tells the global tech, finance, stock market, media, economy, lifestyle, real estate, AI and innovative stories you want to know.
Read more »
Kim observes tests of North Korea's new reconnaissance and attack dronesNorth Korea says leader Kim Jong Un observed tests of newly developed reconnaissance and attack drones this week and called for their increased production. Photos released by state media Thursday showed Kim observing what appeared to be a large reconnaissance drone roughly resembling Boeing’s E-7 Wedgetail airborne warning and control aircraft.
Read more »
Kim observes tests of North Korea’s new reconnaissance and attack dronesThe Associated Press
Read more »
Verizon unveils new trailer to command and control first responders dronesSebastian, a veteran of a tech writer with over 15 years of experience in media and marketing, blends his lifelong fascination with writing and technology to provide valuable insights into the realm of mobile devices.
Read more »
2025 Morgan Supersport | PH ReviewNew look, new underpinnings, new mission - is the Supersport truly a new dawn for Morgan?
Read more »