Forthcoming Automatica paper: Multiplexed MPC

Keck Voon Ling, Jan Maciejowski, Arthur Richards, Bing Fang Wu, Multiplexed model predictive control, Automatica, Available online 15 December 2011, ISSN 0005-1098, 10.1016/j.automatica.2011.11.001.

Abstract: This paper proposes a form of MPC in which the control variables are moved asynchronously. This contrasts with most MIMO control schemes, which assume that all variables are updated simultaneously. MPC outperforms other control strategies through its ability to deal with constraints. This requires on-line optimization, hence computational complexity can become an issue when applying MPC to complex systems with fast response times. The Multiplexed MPC (MMPC) scheme described in this paper solves the MPC problem for each subsystem sequentially, and updates subsystem controls as soon as the solution is available, thus distributing the control moves over a complete update cycle. The resulting computational speed-up allows faster response to disturbances, which may result in improved performance, despite finding sub-optimal solutions to the original problem. This paper describes nominal and robust MMPC, states some stability results, and demonstrates the effectiveness of MMPC through two examples.

Keywords: Predictive control; Distributed control; Multivariable control; Periodic systems; Constrained control

Recruiting again! PhD opportunity on Micro Air Vehicle navigation

UPDATE: this post has been filled.

A fully-funded studentship is available for a PhD on the topic of autonomous indoor flight of a Micro Air Vehicle (MAV) using computer vision. The focus of the project is on the guidance, sensing and control aspects of the problem. It will build on prior work on visual SLAM and MAV control, and off-the-shelf platforms will be used for experiments. In particular, the project will seek to optimize trajectories and motions for the use of vision as a navigation sensor. The work will combine analysis, numerical simulation and hardware experiments. The studentship is supported by dstl (Defence Science and Technology Laboratory: http://www.dstl.gov.uk ) and is part of a larger activity to develop future capabilities for building exploration in hazardous environments.

The studentship will be held in the Department of Aerospace Engineering, part of Bristol’s Faculty of Engineering. The project will be run in collaboration with the Department of Computer Science and will be affiliated with the Bristol Robotics Lab, utilizing the large indoor flying arena for experiments. Stipends will be at the standard level of £13,590 p.a. plus a top-up of £3,000 p.a. The project is aiming to begin in early 2012.

The successful applicants will all hold good (minimum 2:1, preferable first class) first degrees in engineering, robotics, computer science, mathematics or a related discipline. Experience in one or more of: computer vision; flight mechanics; control; or numerical optimization would be highly desirable. Due to funding restrictions, only EU nationals are eligible for this studentship.

For more information, please see the pages of the academic supervising team, listed below. For general enquiries, please contact Arthur Richards, whose contact details can be found on the page below. To apply, visit http://www.bristol.ac.uk/prospectus/postgraduate/2012/apply.html . On your application form, identify “PhD in Aerospace Engineering” as your chosen degree programme, “Flying with Vision” as your desired topic and Dr Arthur Richards as your proposed supervisor. Please use the “Proposed Research” section to discuss your interest in this particular topic.

- Andrew Calway: http://www.bris.ac.uk/engineering/people/person/andrew-d-calway/overview.html
- Walterio Mayol-Cuevas: http://www.bris.ac.uk/engineering/people/person/walterio-w-mayol-cuevas/overview.html
- Arthur Richards: http://www.bris.ac.uk/engineering/people/arthur-g-richards/overview.html
- Tom Richardson: http://www.bris.ac.uk/engineering/people/tom-s-richardson/overview.html

EUCASS

Hermitage, St Petersburg

I'm just back from attending the European Conference on Aerospace Sciences (EUCASS) in St Petersburg.  Visiting Russia and the imposing city of St Petersburg was a great experience, but I'll try to keep to the technical side in this post.

There was much coverage of launchers, not really my field but interesting anyway.  A concept caught my eye for using a GlobalHawk UAV to lift a launcher rocket for small satellites, similar to the Pegasus system - only got a brief glimpse though.

My student Colin Greatwood presented his work on collision alerting for helicopters, while colleague Massimiliano Saponara from Thales Alenia Space Italy presented results from our ORCSAT study with ESA.  It was interesting as well to see Princeton Satellite Systems proposing something MPC-like for rendezvous in their Space Rapid Transit system.

Finally, an idea: has anyone ever tried MPC for aircraft fuel system management?

GENSPACE: air traffic control for beginners

Photo courtesy EUROCONTROL

I'm just back from two days at the EUROCONTROL Experimental Centre just outside Paris, participating in the GENSPACE course for air traffic control experience.  Aimed at researchers and others in ATC-related activities, this is a chance to gain a bit of first-hand experience in ATC through presentations and simulations.  The EEC has extensive facilities including 44 controller consoles (see photo), 35 pilot consoles, all networked together for simulations of up to 350 aircraft.  The GENSPACE experience includes sessions controlling en-route, TMA and approach sectors, guided by experienced instructors.  I certainly learnt a lot.  It was interesting to see the different conflict resolution strategies in different sector types.  Even at our easy-paced simulations, with a group of about 16 of us covering an area over Vienna, it was pretty draining to manage an hour of traffic.  The experience will prove invaluable for our two forthcoming ATC projects with SESAR.  Thanks to EUROCONTROL and the HALA research network for the opportunity.

New journal paper: airport taxi optimization

UPDATE: now in print

Clare, G. L.; Richards, A. G.; , "Optimization of Taxiway Routing and Runway Scheduling," Intelligent Transportation Systems, IEEE Transactions on , vol.12, no.4, pp.1000-1013, Dec. 2011, doi: 10.1109/TITS.2011.2131650

Abstract: This paper describes a mixed-integer linear programming optimization method for the coupled problems of airport taxiway routing and runway scheduling. The receding-horizon formulation and the use of iteration in the avoidance constraints allows the scalability of the baseline algorithm presented, with examples based on Heathrow Airport, London, U.K., which contains up to 240 aircraft. The results show that average taxi times can be reduced by half, compared with the first-come–first-served approach. The main advantage is shown with the departure aircraft flow. Comparative testing demonstrates that iteration reduces the computational demand of the required separation constraints while introducing no loss in performance.

ORCSAT, or How to Catch a Basketball in Mars Orbit

Recently we have been working on a European Space Agency (ESA) project called "On-line Reconfiguration Control System and Avionics Technologies" or "ORCSAT".  Our role, in collaboration with the control group at the University of Cambridge, has been to design a Model Predictive Control (MPC) system for spacecraft rendezvous.  The scenario is the rendezvous in Mars orbit with a sample canister, launched from the Martian surface and to be transported back to Earth for analysis.  Loosely speaking, we're trying to catch a basketball from 50km away...

MPC works by using a trajectory optimizer to design the manoeuvres in real time to use minimum propellant.  The challenge is to design the optimizer to balance complexity with speed.  Too detailed, and the optimization takes too long to solve: too approximate, and the manoeuvre doesn't go where predicted and you burn too much fuel.  The figure on the right shows an example of the resulting trajectory.  Bristol researcher Paul Trodden developed the far-term MPC that gets us to a similar orbit to the target and close enough to track it.  Then Cambridge researcher Ed Hartley's controllers take over to "hop" closer to the target and finally intercept it.  Initial results suggest that MPC is very efficient, reducing the fuel needed to accomplish the rendezvous.

Our collaborator Alberto Bemporad from the University of Trento developed MPC software to enable all these controllers to share the same optimizer.  Also in the consortium are GMV in Spain, who provided a rendezvous simulator, Reliacon of the Netherlands, who are verifying the controllers, and Thales Alenia Space Italy (TAS-I), who lead the project.  TAS-I and GMV are now implementing and testing the MPCs as the project continues.