Agilicious: Autonomous drone can dodge obstacles at breakneck 40 mph speeds
Autonomous drones can play a crucial role in traffic management, emergency response, border surveillance, and delivery services in the future. To advance the capabilities of autonomous uncrewed aerial vehicles (UAVs), scientists and engineers are working on various solutions. Recently, a team of researchers has developed an open-source hardware and software model that promises to improve the agility of quadrotors (drones that run on a four-rotor system).
The model is named ‘Agilicious’, and the researchers claim that it can make quadrotors better suited to avoid obstacles and track trajectories even at high speeds during their journeys. Moreover, Agilicious is useful in both actual flights and virtual simulations. “By open-sourcing Agilicious, we provide the research and industrial community access to a highly agile, versatile, and extendable quadrotor platform,” said the researchers.
How does Agilicious improve the flight capabilities of a quadcopter?
Although quadrotors are highly agile UAVs, but their agility needs to be improved further to employ them for time-sensitive activities such as rescue missions, delivery operations, and shipment inspection. These drones often struggle when it comes to adaptive control and tracking trajectories. To overcome these limitations, the researchers from the University of Zurich decided to make a platform that uses multiple software and hardware tools to enhance a quadrotor's agility.
Under their platform Agilicious, they employed advanced onboard vision sensors, flight monitoring systems, and various other components that eventually improved the drones' real-time perception, trajectory tracking, and flight planning skills. Moreover, the platform also features NVIDIA Jetson TX2, a powerful and energy-saving AI module that performs complex computational tasks to support its hardware functions.
The researchers then conducted various experiments to test the platform’s impact on the UAV’s performance. Surprisingly, even at high speeds ranging between 30 to 43 mph (50 to 70 kph), the autonomous drones with Agilicious were able to perform motion-capture trajectory tracking. Later, the platform was also tested for virtual reality simulations.
During both types of tests, researchers observed an improvement in the quadrotors' real-time response and obstacle-avoiding skills.
What does it mean for the future of drones?
Excited with the successful testings of Agilicious, the authors wrote in their study, “Agilicious offers a unique quadrotor testbed to accelerate current and future research on the fast autonomous flight. Its versatility in both hardware and software allows deployment in a wide variety of tasks, ranging from exploration or search and rescue missions to acrobatic flight.”
The researchers believe that this open-source framework brings us closer to scaling up the use of drones for real-world applications. Moreover, it also paves the path for research on other types of autonomous robot technology. The advanced computing capabilities and efficient hardware components make Agilicious a perfect platform for autonomous UAVs' current and future generations.
The study concerning Agilicious is conducted by researcher Philipp Foehn et al. and is published in the journal Science Robotics.