Two papers from the Intelligent Space Robotics lab. (Skoltech Centre for Digital Engineering) have been accepted to the 2025 International Conference on Unmanned Aircraft Systems, IEEE ICUAS 2025, will take place in Charlotte, NC, USA, on May 14-17.

Roohan Khan, MSc-2: Engineering Systems at the Intelligent Space Robotics lab (Skoltech Center for Digital Engineering)

"AgilePilot: A drone agent with real-time motion planning based on deep reinforcement learning (DRL) and object detection for dynamic environments"

"The AgilePilot system helps UAVs fly quickly and safely among moving obstacles and changing targets. Conventional methods often fail to cope with sudden changes in the environment, but AgilePilot uses artificial intelligence and computer vision to enable UAVs to analyse the environment in real time, predict object movement and adjust routes in a timely manner."

The article is publicly available at the link.

A video demonstration of the AgilePilot project is here.

Malaika Zafar, MSc-2: Engineering Systems at the Intelligent Space Robotics lab (Skoltech Center for Digital Engineering)

"HetSwarm is a system for collaborative navigation of a heterogeneous swarm (UAV + ground robot) in dynamic and dense environments. The UAV uses a path planner based on an artificial potential field, adapting the trajectory in real time, while the ground robot follows its trajectory, maintaining connectivity through impedance links. Additionally , there are impedance links between UGV and low height obstacles that help it to avoid collisions with them since they do not interfere with the drone. Tests showed 90% success rates in 30 scenarios, with the UGV exhibited an average deviation of 45 cm near obstacles. Simulations in Gym PyBullet confirmed the system's robustness for real-world tasks in challenging environments."

The article is publicly available at the link.

A video demonstration of the HetSwarm project is here.