Autonomous Drones Landing on Moving Vehicles

Autonomous and precise take-off and landing are crucial for various applications involving drones. For instance, docking into recharging stations, or in information gathering and delivery applications, drones need to reach precise positions and return to a base accurately. However, the challenge becomes more significant when it comes to landing on a moving vehicle. This task requires even more advanced capabilities, especially if the drone lacks a precise localization system. Additionally, the operational time is often limited in such scenarios, making it essential for the drone to cooperate with another vehicle.

By enabling vertical take-off and landing on a moving vehicle, UAV systems can be deployed from boats, trains or cars near target locations of the mission. This approach allows for short-term flights that effectively utilize the capabilities of UAVs while extending their operational range when combined with a moving vehicle.

Fly4Future’s founder & CTO Dr. Martin Saska and some of our employees were core members of the MRS CTU international team competing at the renowned Mohamed Bin Zayed International Robotics Challenge (MBZIRC), March 16-18, 2017, at the Yas Marina Circuit in Abu Dhabi, UAE. Read more about the competition here.

Under the leadership of Dr. Saska, the team made a significant contribution and achieved the silver position in this prestigious competition. Their focus was on the category of autonomous landing on a moving vehicle, the area that has been described in hundreds of academic works but rarely performed in the real-world conditions. The challenge aimed to autonomously localize a moving vehicle in the competition arena by a single MAV and then land on a landing platform carried by the vehicle.

To ensure stability during landing, magnetic legs are used to attach themselves securely to the vehicle. This entire process is fully autonomous, from take off to landing.  Moreover, a novel control system uses a Model Predictive Controller (MPC) in real-time to generate a reference trajectory for the drone. This trajectory is then tracked by a nonlinear feedback controller, which helps minimize any position errors between the drone and its predicted target. In conclusion, this combination of technologies enables accurate tracking of vehicle motion predictions with minimal mis-positioning. Consequently, this efficient solution achieved the fastest landing among all competing teams at the MBZIRC 2017.

The members of the successful MRS CTU competition team were later recruited by Fly4Future and have since become valued employees who, thanks to the experience gained in the competition, expanded the expertise of the company. Currently, Fly4Future is actively involved in the process of commercializing the results achieved during the competition.