University of Hong Kong Engineers Unveil SUPER: The Agile, High-Speed Autonomous Aerial Robot
Engineers at the University of Hong Kong have unveiled an innovative aerial robot, dubbed SUPER, designed for high-speed navigation in uncharted territories. This advanced drone is set to revolutionize the field of autonomous Micro Air Vehicles (MAVs) with its sophisticated technology and wide-ranging applications.
SUPER’s Revolutionary 3D LiDAR Sensor
At the heart of SUPER’s capabilities lies a lightweight 3D LiDAR sensor, which provides precise and long-range detection of obstacles. This sensor ensures that the drone can identify and avoid obstacles while maintaining high speeds. According to the research team, SUPER can fly over 20 meters per second, crossing complex obstacle courses without any incidents.
The drone uses an efficient planning framework to generate flight trajectories directly from the LiDAR point clouds. In each re-planning cycle, the system generates two trajectories: a safe one for known free spaces and a faster one for all spaces. This dual approach has proven to be highly effective, reducing failure rates by 35.9 times and decreasing planning time by half.
“SUPER represents a significant milestone in autonomous MAV systems, marking the transition from laboratory research to real-world applications,” stated the researchers in their study abstract.
Agile MAV Navigation
Birds are renowned for their ability to soar through cluttered spaces at high speeds with minimal mishaps. MAVs aim to emulate this agility for practical applications such as search and rescue operations. However, achieving safe, high-speed flight in unknown environments requires a delicate balance between agility, long-range sensing, and efficient trajectory planning, all done on-board.
Many previous MAV projects focused on speed using pre-calculated paths or reinforcement learning methods. However, these approaches often required predefined environments or external sensors, limiting their real-world applicability. The researchers at the University of Hong Kong found that their dual-trajectory technique, though it introduced some computational delays, significantly increased speed while ensuring safety.
While most MAVs use vision sensors for navigation, these sensors are susceptible to motion blur, have limited range, and struggle in low light conditions. In contrast, SUPER utilizes a lightweight 3D LiDAR sensor, allowing it to operate effectively in various settings, including low light.
Field tests demonstrated that SUPER could navigate through a forest, successfully avoiding trees and branches, and even follow a moving target, such as a person, without any issues.
Efficient Drone Autonomy
SUPER’s reliance on 3D LiDAR for navigation provides significant advantages over vision-based systems in terms of range, accuracy, and reliability, especially in low light conditions. Traditional LiDAR-equipped drones have been limited by bulky and expensive sensors that impede efficiency and speed.
To overcome these limitations, the University of Hong Kong team designed SUPER with a compact, high-thrust-to-weight ratio structure, incorporating a lightweight Livox MID360 LiDAR sensor. This design allowed the drone to maintain high speeds while remaining agile and maneuverable.
The drone’s trajectory planning is optimized using gradient-based methods, ensuring higher success rates compared to traditional mixed-integer quadratic programming. SUPER’s dual-trajectory approach further enhances safety by maintaining a backup path and dynamically adjusting switching times to optimize speed.
Thanks to onboard LiDAR-inertial odometry, SUPER can operate independently in environments without access to Global Navigation Satellite Systems (GNSS). This feature makes it versatile for a variety of real-world applications, from exploration and logistics to inspection and search-and-rescue.
The future of SUPER lies in further technological advancements. Smaller, lighter LiDAR sensors with extended ranges could boost the drone’s agility and speed, enabling it to navigate even tighter spaces. Enhanced aerodynamics and improved motion prediction techniques would also contribute to greater efficiency.
The research team envisions SUPER as an ideal solution for operations in diverse, unstructured environments, where low computational demands and high-speed safety are critical.
Details of the study can be found in the journal Science Robotics.
Conclusion
The University of Hong Kong’s development of SUPER represents a significant leap forward in the capabilities of autonomous aerial vehicles. With its advanced 3D LiDAR sensor, efficient trajectory planning, and robust safety systems, SUPER sets the stage for a new era of practical and reliable MAV applications.
As technology continues to evolve, we can anticipate further innovations that will push the boundaries of what MAVs can achieve. The potential uses of drones like SUPER in search and rescue, logistics, inspection, and exploration are vast, and their development underscores the importance of continued research in this field.
Stay tuned for more updates and advancements in this exciting area of technology. We invite you to share your thoughts and questions in the comments section below.
Thank you for reading. If you enjoyed this article, consider subscribing to our newsletter to receive the latest updates and insights from the world of innovation.
Don’t forget to share this article on your social media channels to spread the word about these groundbreaking advancements.
Join the conversation:
Leave a comment |
Subscribe to our newsletter |
Share on Facebook |
Share on Twitter