Autonomous Sailing

aus MariaTheresia, der freien Wissensdatenbank

Org.Unit	DMU/PhD
(Schul)jahr	2004ff.
Teilnehmer	Roland Stelzer (mailto:admin@roland-stelzer.info)
Keywords	autonomous sailing, route optimisation, weather routing

Working Title: Real-Time Route Optimisation for Sailboats Using Artificial Intelligence

Es handelt sich hier um eine laufende Dissertation am Centre for Computational Intelligence (http://www.cci.cse.dmu.ac.uk/cci/) and der De Montfort University in Leicester (http://www.dmu.ac.uk). Es ist jeder eingeladen mitzuarbeiten bzw. seine Ideen einzubringen. Besonders im Bereich der HW-technischen Umsetzung (Sensoren/Aktoren am Modellboot, Wetter-Sensorbojen, Ortungssystem, ...) bin ich für jede Unterstützung dankbar!

Siehe auch Microtransat: Wettbewerb für autonome Segelboote

Inhaltsverzeichnis

1 Main Aims of Investigation 2 Relation to Previous Published Work 3 Proposed Methods of Investigation 4 Anticipated Outcomes 5 Key References

Main Aims of Investigation

• To research in the field of autonomous sailboat navigation
• To develop a novel methodology for real time route optimisation for sailboats in dynamic and uncertain environments
• To explore and analyse the most appropriate artificial intelligence techniques for this task
• To simulate the route optimisation on a computer
• To verify the algorithm using a wheeled robot and/or a yacht model

Relation to Previous Published Work

The proposed research project combines artificial intelligence (AI), sailing techniques, and route optimisation. Even though research has been carried out in each of these areas, the combination of all three has not been investigated yet and represents a challenging topic. The most promising initiative in the field of AI-based sailing aids is the RoboSail project [17]. The ultimate goal within RoboSail is "to create a semiautonomous, intelligent, computer system, which can learn to steer a sailboat optimally, in close cooperation with the human sailor onboard" [17]. The authors develop a multi-layer robot control architecture, which is based on a merge of the Subsumption Architecture by Brooks [4] and the Xavier Architecture by Simmons [13]. Machine learning has been used in a knowledge discovery process to build up a rule base that describes knowledge about the activity of sailing [18]. However, the aspect of routing strategies was not covered by this rule base. On the other hand, there exist various publications about AI-based route optimisation, mainly based on evolutionary algorithms [11,16,14] or fuzzy logic [20]. None of these AI-based approaches addresses the special situation of sailing boats in a dynamic environment. Some papers deal with ship routing, but with conventional, non-AI approaches [2,3,7,8,10]. The proposed research project will merge two existing fields of AI activity and will fill the gap in between sailboat control and route optimisation.

Proposed Methods of Investigation

A central part of the research will be a modular simulation framework to compare different approaches to route optimisation. Modules of the system are an electronic map (static environment), time-dependent weather data (dynamic environment), and the real-time route optimisation algorithm. Real-time in this context means, that the system reacts on changes in the dynamic environment and adapts the route if this is necessary. Data interfaces allow that each of the modules can be changed or replaced. The ultimate goal is to find a routing algorithm, which determines the optimal route for all possible static and dynamic environments. "A sailing vessel resides in a complex, dynamical environment, governed by aerodynamics and hydrodynamics, two mathematically intractable problems." [17] Therefore, research within the proposed project will mainly focus on soft computing techniques to find a robust, almost-optimal solution.

The correlation between weather conditions and boat speed has to be defined. Experts’ knowledge and literature about the physics of sailing will be used to configure a system based on AI representing this relationship.

A wheeled robot and/or a yacht model will be used to verify the results of the computer simulation. Important aspects in this context are a multi-layer robot control architecture, identification of important sensors and actuators, low level control to carry out manoeuvres, and communication between the components of the system (robot control architecture, sensors and actuators, routing system). Real-time weather data can either be gathered from sensor buoys, or from meteorological institutes.

Anticipated Outcomes

• Contribution to theories of AI based route optimisation
• Improved high level navigation aid for sailing vessels
• Computer simulation to demonstrate the algorithm
• Wheeled robot and/or yacht model and related infrastructure to prove the feasibility
• Publications at conferences and/or in scientific journals

Key References

• [1] P. Adriaans, "From knowledge based to skill-based systems: Sailing as a machine learning challenge," in Machine Learning: EMCL 2003: 14th European Conference on Machine Learning (L. T. Mada Lavrac, Dragan Gamberger, ed.), vol. 2837/2003, pp. 1-8, University of Amsterdam, Springer-Verlag Heidelberg, September 2003. ISSN: 0302-9743, ISBN: 3-540-20121-1, DOI: 10.1007/b13633.
• [2] T. Allsopp, A. J. Mason, and A. B. Philpott, "Optimal sailing routes with uncertian weather," in Proceedings of the 35th Annual Conference of the Operations Research Society of New Zealand, pp. 65-74, 2000.
• [3] T. Allsopp, A. Mason, and A. Philpott, “Optimising yacht routes under uncertainty,” in Proceedings of the 2000 Fall National Conference of the Operations Research Society of Japan, (Japan, Tokyo), pp. 176–183, September 2000.
• [4] R. A. Brooks, “A robust layered control system for a mobile robot,” IEEE Journal of Robotics and Automation, vol. 2, pp. 14–23, September 1985.
• [5] D. Gallardo, O. Colomina, F. Florez, and R. Rizo, “A genetic algorithm for robust motion planning [online],” tech. rep., Universidad de Alicante, San Vicente, Spain, Available from www.rvg.ua.es/publications/gallardo98b-IEAAIE.pdf [Accessed 6 October 2004], 1998.
• [6] Y. Hayashi, N. Wakabayashi, T. Nanri, and H. Wake, “Database concept for static navigational information,” in Proceedings of the IEEE 1996 Position Location and Navigation Symposium, (Atlanta, USA), pp. 96–102, April 1994.
• [7] J. Hinnenthal, “Optimierung von schiffsrouten auf grundlage numerischer simulation (route optimization based on numerical simulations),” in Sommertagung der Schiffbautechnischen Gesellschaft, (Wismar), Technical University of Berlin, June 2003.
• [8] J. Hinnenthal and S. Harries, “A systematic study on posing and solving the problem of pareto optimal ship routing,” in COMPIT’04, (Siguenza, Spain), May 2004.
• [9] D. Klasing, Seemannschaft - Handbuch fr den Yachtsport. ramon red. gliewe ed., December 2003. ISBN: 3768805239.
• [10] H. Lee, G. Kong, S. Kim, C. Kim, and J. Lee, “Optimum ship routing and it’s implementation on the web,” in First International Workshop on Advanced Internet Services and Applications, pp. 125–136, 2002. ISBN:3-540-43968-4.
• [11] D. K. Liu, H. Lau, and G. Dissanayake, “A hierarchical approach to a multilevel genetic algorithm for vehicle path planning,” in The Second International Conference on Computational Intelligence, Robotics and Autonomous Systems (CIRAS 2003), Faculty of Engineering, University of Technology, Sydney, Australia, December 2003. PS06-4-01.
• [12] A. J. Mason and A. B. Philpott, “Optimising yacht routes under uncertainty,” in 15th Chesapeake Sailing Yacht Symposium, pp. 89–98, 2001.
• [13] R. Simmons, R. Goodwin, K. Haigh, S. Koenig, and J. Sullivan, “A layered architecture for office delivery robots,” in Proceedings of the First International Conference on Autonomous Agents (Agents’97) (A. L. A. for Office Delivery Robots, ed.), (New York), pp. 235–242, ACM Press, February 1997. ISBN 0-89791-877-0.
• [14] R. Smierzchalski and Z. Michalewicz, “Adaptive modeling of a ship trajectory in collision situations at sea,” in 2nd IEEE World Congress On Computational Intelligence, ICEC’98, (Alaska, USA), 1998.
• [15] A. Treby, “Optimal weather routing using ensemble weather forecasts,” in 37th Annual ORSNZ Conference, (University of Auckland), November 2002.
• [16] J. Tyler, L. Booker, G. Brisbols, and B. Canova, “Route planning for individual combatants using genetic algorithms,” in In Proceedings of the Sprint Simulation Interoperability Workshop, pp. 177–185, The MITRE Corporation, May 1997.
• [17] M. L. van Aartrijk, C. P. Tagliola, and P. W. Adriaans, “Ai on the ocean: the robosail project,” in European Conference on Artificial Intelligence 2002 (F. van Harmelen, ed.), pp. 653–657, RoboSail systems BV, IOS Press, 2002.
• [18] M. van Aartrijk and J. Samoocha, “Learning to sail - knowledge discovery in embedded adaptive systems,” in Proceedings of European Symposium on Intelligent Technologies, Hybrid Systems and their implementation on Smart Adaptive Systems (EUNITE 2003), RoboSail systems BV and Perot Systems BV, Verlag Mainz, Aachen, 2003.
• [19] W. H. Warden, “A control system model for autonomous sailboat navigation,” in IEEE Proceedings of Southeastcon, vol. 2, (Williamsburg, VA , USA), pp. 944–947, School of Industry and System Engineering, Georgia Inst. of Technol., Atlanta, GA, July - October 1991. IEEE Catalog Number: 91CH2998-3.
• [20] P. Webb, C. Fayad, and C. Breitenbach, “The integration of an optimised fuzzy logic navigation algorithm into a semi autonomous robot control systen,” in The International Workshop on Recent Advances in Mobile Robots, (Leicester, UK), June 2000.
• [21] J. Wolfe, “The physics of sailing [online],” tech. rep., School of Physics, The University of New South Wales, Available from: http://www.phys.unsw.edu.au/~jw/sailing.html [Accessed 5 October 2004].