Research

Flight dynamics & control · Guidance · Unmanned and autonomous systems · AI, machine vision & robotics

Our research sits at the intersection of classical flight mechanics and modern artificial intelligence. Building on expertise that spans aerospace guidance and control, computer and machine vision, and industrial robotics, we develop methods that allow aerospace vehicles to perceive, decide and act autonomously, from the guidance law of a missile to the robotic arm that refuels an aircraft.

Outlook

Looking ahead, our aim is to carry learning-enabled autonomy from simulation into dependable flight. Our research programme is organised around four directions.

01

Trustworthy learning-enabled control

Combining reinforcement learning with robust and optimal control so that learned guidance and control policies come with guarantees of stability, constraint satisfaction and predictable behaviour.

02

Simulation-to-reality transfer

Photorealistic digital twins, synthetic data and hardware-in-the-loop testing to close the gap between training environments and real aircraft, sensors and actuators.

03

Cooperative and close-proximity autonomy

Swarms, formation flight, manned–unmanned teaming and autonomous refuelling, in which several vehicles must perceive, predict and coordinate with one another in real time.

04

Intelligent operations and maintenance

AI-assisted robotic inspection, refuelling and ground handling that make aircraft operations faster, safer and less dependent on manual work.

Methods & Tools

  • MATLAB / Simulink
  • Python
  • PyTorch
  • TensorFlow
  • ROS / Gazebo
  • NVIDIA Isaac Sim / Isaac Lab
  • SolidWorks
  • Blender
  • MVTec HALCON
  • OpenCV
  • XFLR5
  • AVL
  • Tornado
  • FlightGear
  • Digital DATCOM
  • RASAero
  • Ansys Electronics
  • Altair Feko

Research Themes

Flight Dynamics & Nonlinear Simulation

High-fidelity six-degree-of-freedom models of aircraft, rotorcraft, missiles and rockets, and nonlinear simulation environments for analysing stability, performance and handling qualities. We also develop probabilistic models of pilot behaviour to capture the human element in manned and optionally piloted flight.

6-DoF modelling · nonlinear simulation · pilot modelling

Guidance & Control

Guidance laws and flight control systems for precision munitions and air vehicles, including impact-angle-constrained terminal guidance, integrated guidance and control, high-angle-of-attack missile control, and robust and model predictive control.

missile guidance · H∞ & MPC · robust control

Learning-Based Guidance & Control

Deep reinforcement learning agents that learn guidance and control policies directly from interaction, with staged training schemes that improve generalisation across wide engagement envelopes. Applications include agile-target interception, high-angle-of-attack missile control and the autonomous vertical landing of rockets.

deep RL · integrated G&C · generalisation

Trajectory Optimisation

Constrained trajectory optimisation for missiles and air vehicles, comparing and combining optimal-control, evolutionary and learning-based approaches to satisfy terminal, path and actuator constraints in real time.

optimal control · genetic algorithms · AI-driven optimisation

Unmanned, Autonomous & Cooperative Systems

Autonomy for unmanned aerial, ground and maritime vehicle systems: path planning and target tracking, swarm intelligence and the cooperative guidance of multiple agents, and close-proximity flight such as formation flight and autonomous air-to-air refuelling.

swarms · path planning · formation flight

Machine Vision & Robotics for Aerospace

Vision-based perception and robotic manipulation for aerospace operations: learning-based pose estimation trained on photorealistic synthetic data and digital twins, autonomous ground and aerial refuelling, and thermal-imaging-based robotic inspection of aircraft structures.

pose estimation · synthetic data · robotic inspection