Ashley Foster presents Multi-UAV Coverage Path Planning for Wind Farm Inspection
On 2026-02-13 - 2026-02-13 11:00:00 at E126
Speaker: Ashley Foster, PhD student at the University of Plymouth, with a
research focus on multi-robot coverage planning for offshore renewable
operations and maintenance.
Abstract:
This talk presents Reinforcement Learning-enhanced Coalition-Based
Metaheuristics (RL-CBM), a framework for online, distributed task allocation in
Multi-Robot Coverage Path Planning. RL-CBM integrates reinforcement learning
into a coalition-based metaheuristic to adaptively select diversification and
intensification operators during execution, enabling robots to learn effective
planning strategies as a team. The approach is demonstrated on an online UAV
wind turbine inspection problem, where robots must cope with uncertainties such
as communication loss and robot failures. Through comparative evaluation against
a greedy online coverage strategy, RL-CBM achieves reduced makespan and
increased robustness. The talk will primarily focus on the details of the
algorithm and experimental evaluation of RL-CBM in multi-robot scenarios.
research focus on multi-robot coverage planning for offshore renewable
operations and maintenance.
Abstract:
This talk presents Reinforcement Learning-enhanced Coalition-Based
Metaheuristics (RL-CBM), a framework for online, distributed task allocation in
Multi-Robot Coverage Path Planning. RL-CBM integrates reinforcement learning
into a coalition-based metaheuristic to adaptively select diversification and
intensification operators during execution, enabling robots to learn effective
planning strategies as a team. The approach is demonstrated on an online UAV
wind turbine inspection problem, where robots must cope with uncertainties such
as communication loss and robot failures. Through comparative evaluation against
a greedy online coverage strategy, RL-CBM achieves reduced makespan and
increased robustness. The talk will primarily focus on the details of the
algorithm and experimental evaluation of RL-CBM in multi-robot scenarios.