Competition

The DRAFT team is currently gearing up the challenge as they participate in the prestigious Leonardo Drone Contest 2023. Organized by Leonardo in collaboration with six leading Italian universities, this Open Innovation competition aims to drive advancements in Artificial Intelligence applied to Uncrewed Systems. The project seeks to foster an innovative ‘ecosystem,’ uniting the strengths of large corporations, universities, SMEs, and national startups. 

Over the course of the years, the students have been diligently crafting an autonomous drone guidance system, leveraging the combined resources and expertise of their universities and Leonardo. For the upcoming competition, each of the six PhD students has assembled a diverse, multidisciplinary team drawn from various engineering disciplines and academic levels. This collaborative effort not only enhances individual skill sets but also aligns seamlessly with the contest’s overarching objectives. Spanning three years and featuring annual events, a scientific symposium, and a thrilling competition, the Leonardo Drone Contest has evolved into a beacon of innovation since its launch in 2019. 

The previous editions witnessed fierce competition among the six university teams, testing their mettle on the ‘airfield’ at Leonardo’s Turin site. With complexity levels escalating each year, the contest has evolved into a true drone battleground. The culmination of this experimental cycle is slated for October 6 and 7, 2022, offering the final showdown for the six teams and their last chance to claim ultimate victory before embarking on new adventures in the realm of drone technology. 

With unwavering commitment, the DRAFT Polito is set to return for the highly anticipated 2023 installment of the competition, demonstrating their enduring passion for advancing the frontiers of autonomous drone technology. 

DFT22 Agares

DFT22 Agares

This is the last evolution of our autonomous drone line-up. Once again, it is designed to participate in the Leonardo Drone Contest 2022 – third edition. In this edition, the drone will operate in an urban environment with a higher level of flexibility regarding the type of task required. The whole mission will be managed autonomously without being controlled by a human pilot. All the necessary computing must be executed on board.

To achieve the mission objectives, the drone uses a complex software suite (developed inside the team) to perform SLAM (Simultaneous Localization and Mapping) attitude determination and computer vision to perceive the environment. The drone only uses cameras and proximity sensors to orient itself since neither LIDAR nor GPS is allowed.

DFT21 Acheronte​

DFT21 Acheronte

This is an autonomous drone we are designing and building with the aim of participating in the Leonardo Drone Contest 2021. The drone is conceived to navigate in a known area representing an urban environment with the purpose of searching for terrestrial bots, reading the Aruco codes to get information about the landing pads to search for in the field, fly toward them by avoiding the obstacles and finally land over. During the whole mission, the drone must perform autonomously, without being controlled by a human pilot. All required computing must be executed on board, thus requiring a low power, high performance computer.

To achieve the mission objectives, the drone uses SLAM (Simultaneous Localization and Mapping) for determining the state of the drone (position, attitude, speed), then computer vision and machine learning help identifying the terrestrial bots and the landing pads, finally obstacle avoidance and motion planning algorithms define the trajectory to follow for reaching the targets. The drone only uses cameras and proximity sensors to orient itself, since neither LIDAR nor GPS are allowed.  

DFT20 Stige

DFT20 Stige

This was the first autonomous drone we designed and built. The aim was to participate in the Leonardo Drone Contest 2020. The drone is conceived to map an unknown area representing an urban environment with the purpose of searching for landing pads in a predefined sequence, fly towards them by avoiding the obstacles and finally land over. During the whole mission, the drone must perform autonomously, without being controlled by a human pilot. All required computing must be executed on board, thus requiring a low power, high performance computer.

To achieve the mission objectives, the drone uses SLAM (Simultaneous Localization and Mapping) for mapping the environment and determining the state of the drone (position, attitude, speed), then computer vision and machine learning help identifying the landing pads, finally obstacle avoidance and motion planning algorithms define the trajectory to follow for reaching the target. The drone only uses cameras and proximity sensors to orient itself and map the environment, since neither LIDAR nor GPS are allowed.