Applications


I am very much interested in practical application of my research. In the following you find some examples of projects I am involved in. I am very open to new domains and new applications, and I'd be happy to start new collaborations.
Please send me an email if you see potential links!

Underwater Robotics

blocks_image This project is funded by the EU FP7 Project PANDORA.

PANDORA is a three-year project that will develop plan-based intelligent control of underwater vehicles operating for long periods without human intervention. A key goal is to significantly reduce the frequency of assistance requests. Following the Deep Water Horizon oilfield disaster in the Gulf of Mexico in 2010, contractors are developinghighly actuated and hover capable autonomous underwater vehicles (AUVs) for subsea inspection and intervention. PANDORA is focussing on automating inspection and maintenance of subsea oilfields.
You can find more details about this application in the following paper:
Planning Inspection Tasks for AUVs.
M. Cashmore, M. Fox, T. Larkworthy, D. Long, D. Magazzeni.
Proceedings of OCEANS'13 MTS/IEEE.

The video below shows an early test from the first year of PANDORA. The Nessie AUV from the Ocean Systems Lab of Heriot-Watt University inspects pipes and pillars in the diver centre at Fort William, Scotland.


Autonomous Robots

blocks_image This project is funded by the EU FP7 Project SQUIRREL.

SQUIRREL addresses clutter in an open world by actively controlling the environment, and incrementally learning to extend the robot’s capabilities while doing so. We term this the B3 (bit by bit) approach, as the robot tackles clutter one bit at a time and continuously extends its knowledge as new bits of information become available. Squirrel is inspired by a user driven scenario, that exhibits all the rich complexity required to convincingly drive research, but allows tractable solutions with high potential for exploitation.
You can find more details about this application on the project website: http://www.squirrel-project.eu/

AUV Mission Control

blocks_image This project is a collaboration with the Monterey Bay Aquarium Research Instititute (MBARI), California.

We consider the problem of tracing the structure of oceanological features using Autonomous Underwater Vehicles (AUVs). Solving this problem requires the construction of a control strategy that will determine the actions for the AUV based on the current state, as measured by on-board sensors and the historic trajectory (including sensed data) of the AUV.
You can find more details about this application in the following papers:
Policy Learning for Autonomous Feature Tracking.
D. Magazzeni, F. Py, M. Fox, D. Long, K. Rajan.
Autonomous Robots.
Plan-based Policy Learning for Autonomous Feature Tracking.
M. Fox, D. Long, D. Magazzeni.
Proceedings of the 22nd International Conference on Automated Planning and Scheduling (ICAPS-12).

Multi-Battery Load Management

blocks_image Efficient use of multiple batteries is a practical problem with wide and growing application. We casted this problem as a planning problem under uncertainty and then we developed a new planning technique based on planning for mixed-discrete-continuous problems and Monte Carlo-based policy learning.

Application of the approach leads to construction of policies that, in simulation, significantly outperform those that are currently in use and the best published solutions to the battery management problem. We achieve solutions that achieve more than 99% efficiency in simulation compared with the theoretical limit and do so with far fewer battery switches than existing policies.
You can find more details about this application in the following paper:
Plan-based Policies for Efficient Multiple Battery Load Management.
M. Fox, D. Long, D. Magazzeni.
Journal of Artificial Intelligence Research (JAIR) 44: 335-382. June 2012.

Unit Commitment Problem

blocks_image Unit Commitment is a fundamental problem in power systems engineering, deciding which generating units to switch on, and when to switch them on, in order to efficiently meet anticipated demand.
It has traditionally been solved as a Mixed Integer Programming (MIP) problem but upcoming changes to the power system drastically increase the MIP solution time.

Using AI Planning affords more flexibility in the time points at which generating units can be scheduled and manipulated, which could lead to a potential decrease in solution time and possible cost saving
You can find more details about this application in the following paper:
Challenge: Modelling Unit Commitment as a Planning Problem
J. Campion, C. Dent, M. Fox, D. Long, D. Magazzeni.
Proceedings of the 23rd International Conference on Automated Planning and Scheduling (ICAPS-13).