Reaching the remotest and most inaccessible places on earth, discovering the secrets of abyssal depths and entering space have only been possible thanks to the development of “intelligent machines” capable of undertaking tasks in environments that are very difficult for human beings. But who is behind the design of such artificial systems that facilitate work and life in society? Dr. Richard J. Duro Fernández, a researcher of the Campus do Mar and coordinator of the Integrated Engineering Group (GII) at the University of A Coruña (UDC), is one of the persons responsible for developing such intelligent systems that are inspired in the cognitive behaviour of living beings. He and his research group comprised of more than 50 persons have received many awards for their discoveries, amongst which are the 2002 & 2006 Comerma Engineer Award and the 2007 González-Llanos Naval Engineering Award. The GII Group has likewise not only progressed in science but has also contributed to bringing science closer to society thanks to its multiple dissemination activities, from conferences to congresses and students visits from schools and high schools, that offer students a firsthand view of activity within a research group.
Question (Q).- Artificial neural networks and cognitive mechanisms may sound totally strange to someone who is not involved with research in this field, however, when we use the term Artificial Intelligence, most readers are able to gauge the theme of your research. Can you explain what you do?
Answer (A).- In this field, we seek to generate behaviours in machines that seem intelligent to humans. There are many approximations to this objective and in our case, we have focussed on the study of different techniques inspired in nature (“bio-inspired”), in order to obtain physical systems, such as robots and machines, that can carry out tasks in hostile environments whilst simultaneously adapting to changing situations.
(Q).- How did your interest for robotics start and when did you decide to specialise in autonomous systems?
(A).- My interest really started with intelligence and how one could create intelligent systems by using approximations inspired in nature, such as artificial evolution. When one works on these issues, one realises that intelligence and cognitive capacities in living beings are closely linked to the individual bodies and to the different sensory and action capacities that they offer. This is what led me to delve into the world of robotics and autonomous systems. They were tools that I could use for studying intelligence from a holistic point of view via the body-brain interaction. Obviously, the Engineering School environment in which I was and the needs of the industries that collaborate with us led us increasingly towards systems (that were more or less intelligent) capable of performing industrial tasks in dynamic and non-structured environments such as Shipyards or shipping of petroleum products, by adequately adapting to changing tasks and environments.
(Q).- What brought you to the University of A Coruña?
(A).- My joining the University of A Coruña about twenty years ago was by chance really and it happened after long stays at San Diego State University and the University of California, San Diego, and after the completion of my PhD degree. I was looking for job opportunities in Galicia at the time, which coincided with the creation of the Higher Polytechnic School (EPS) in Ferrol. Things were difficult at first when both the University of A Coruña (UDC) and the EPS were created since there was a paucity of means. Everything was still precarious and a huge effort was needed to obtain very small results. However, one could savour a pioneering ambience that favoured union and collaboration between people from different disciplines and origins to achieve a critical mass. There was great freedom to try out new things since there was practically no limiting guideline structure to follow. In time, this led to the creation of multidisciplinary and dynamic groups within an environment of collaboration with industry, something quite difficult to imagine at the time in other centres with a longstanding history and tradition and with set operating guidelines in place.
(Q).- The interaction between Humans-Machines is a reality now. How does this contribute to the development of society and why is it important to continue research into Artificial Intelligence?
(A).- It is obvious that our civilisation is moving towards an environment of machines and systems that we need to relate to. Such systems make life much easier and facilitate interaction with increasingly complex structures and processes, which are needed to maintain our present day lifestyles in an ever expanding population. This does not mean that man has to become more and more specialised in order to be able to control and operate these systems in such a technical environment. It seems logical to make the machines interact with humans and do so with greater autonomy thereby freeing humans from such chores and providing time for leisure -a rare commodity today. This is what places research into artificial intelligence and development of autonomous systems very high on the list of objectives.
(Q).- What are the main objectives for developing autonomous systems?
(A).- The principal task of such systems is the production of machines that can act and perform tasks in complex and changing environments, without there being any need for their constant supervision and control by humans. This is however not always possible or appropriate.
In the marine field, it ranges from shipbuilding activity (where the so-called traditional production lines do not exist and workers and machinery need to move to the place where shipbuilding actually takes place) to works carried out on the ocean floor where exploration units need to be deployed for cable maintenance operations and repairs of marine and submarine structures. These are environments where there is need for interaction with other machinery and workers that are constantly evolving, and where units are constantly exposed to possible incidents of which we do not have any forewarning.
(Q).- How do you confer autonomy to these machines?
(A).- In order to provide maximum autonomy to such systems, there is need to simultaneously work on two fronts. On the one hand, automatic design techniques are studied and designed to facilitate the best possible morphologies for these machines that are the most suitable for the tasks they are supposed to carry out and for a specific environment. In fact, the starting point is morphologic intelligence in terms of adaptability such that the more adapted and adequate the body, the simpler and more effective is the control or the “brain” unit. On the other hand, cognitive mechanisms are studied and developed for these machines, either as simple sets of rules or through the use of more complex structures that try to imitate the function of animal brains.
(Q).- Artificial intelligence has proved to have multiple applications from domotic engineering processes and patient monitoring to semantic works. Amongst this range of possibilities, can you tell us about the progress made by your research team on the subject of oil spill detection at sea through airborne or satellite imagery, and can you explain what these systems are all about?
(A).- An important issue in many fields, amongst which is the marine field, is “remote sensing”. Our group has a research line dedicated to developing hyperspectral systems not only from the point of view of sensors but also from the point of view of image processing. Hyperspectral images provide each pixel with a spectrum of hundreds or thousands of bands or values instead of providing a RGB colour value. Such spectral resolution facilitates very detailed discrimination between the many materials and elements present in such images.
The problem here is the huge amount of data generated and their processing. A one megapixel image (a small resolution when one considers the normal resolutions available in today’s little photographic cameras) with one thousand spectral bands can have a size of several Gigabytes – an uncommon size in our daily environment. Our team, in collaboration with the Nonlinear Optics Group of the University of Santiago, has recently developed a third generation transportable light hyperspectrometer which is adequate for mounting on cheaper media than the traditional high flying planes and satellites. Several tools have also been developed for segmenting such images, identifying targets, cover classification, and many other applications.
We are currently working on real time image processing to detect and classify elements in line with changes over time, and are also working on developing a small device which can be easily fitted on very small autonomous planes for exploration and civil uses. Examples of applications within the marine field would be searching for shipwrecked persons or detection and follow-up of oil spills.
(Q).- What are the other contributions in the marine field made by your studies, especially for developing fishing activity?
(A).- The Integrated Engineering group (GII), over the last 10 years, has carried out many projects in these two areas through collaboration with companies from these sectors. An autonomous submarine that can operate at 4000 metres depth, automatic longlining systems, aerodynamic effects on ship superstructures, are amongst some of our most important contributions.
A lot of effort has likewise been put into studying and developing systems that use intelligent techniques for preventing dynamic phenomena such as parametric resonance or broaching, wherein fishing vessels or other type of vessels suffer steep inclination as a result of waves and wind action, respectively. A stability management system has also been designed for fishing vessels in addition to carrying out studies for using LPG or kites to propel fishing vessels. The list continues with design of optimised rudder, competition sailing boats, diverse offshore structure elements, not forgetting the ergonomic optimisation of fish processing plants for new deep sea fishing vessels, studies for reducing noise and vibrations for newly designed tugs, amongst others.
(Q).- What are the steps taken when launching the tools you design on to the market given the fact that a large number of your research applications have a high commercial interest?
(A).- For commercially interesting systems, the steps to be followed for market launch in our case are always quite similar. We are a R&D group that always carries out its activities with and for companies from our environment. It is our partner companies on the project that actually commercialise such research results and therefore transfer is immediate, which is how things should be done.
(Q).- What are the projects that you are currently working on?
(A).- The GII group is carrying out a large number of marine projects and these currently number around 30. They are carried out in collaboration with other universities and companies from our environment, and the Galician Naval Technology Centre. Studies are presently being carried out on parametric resonance and prevention in vessels, with a view to improving their safety and reducing the risk of sinking and other damages. In like manner, research is also underway to prevent possible incidents arising from wave battering, designing systems to help captains to improve perception of navigation conditions and vessel stability at all times, a very important issue especially in little fishing vessels.
Some of our shipbuilding projects are focussed on improving the capacities of shipyards such as those related to organisation and production, and others that seek the introduction of new strategies to make such companies even more efficient from a business management point of view. Then there are projects related to construction and anchoring of offshore structures and their management. On the robotic front, work is being carried out on autonomous and modular robots for their use in shipyards. Robots are also being developed for specific uses such as to clean non-magnetic hulls, amongst others. There are several collaboration marine projects being undertaken with ITCs both for improving range of different signals cover in coastal environments by using new buoy designs and for developing systems to detect man overboard.
(Q).- Do these projects have an international projection?
(A).- On the international front, besides the activities carried out by the group with companies, which number more than 100 in total, and which end up in the international market, the GII group is also connected to many other research groups and centres all over the world. The above collaboration has been in the form of joint projects and research stays under the framework of transnational initiatives.
Some such examples are KEDRI (Centre for Knowledge Engineering) at the University of Auckland; The Automation and Control Group for Ship’s Dynamics at the Technical University of Denmark; the Centre for Ships and Ocean Structures (CeSOS) at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway; the LabOceano – Coppe at the Federal University of Rio de Janeiro in Brazil; and the Naval Engineering group at the Technical University of Lisbon. Also relevant are relations maintained through joint projects with the Von Karman Institute of Fluid Dynamics in Brussels, amongst others.
(Q).- The Integrated Engineering Group (GII) brings together the areas of Industrial Organisation (GOI), Naval Engineering (GSA) and Fluid Engineering (GIF). This is a multidisciplinary group that carries out research in wide ranging fields within engineering and computation. What are the benefits of working together?
(A).-To be honest, although this is a single group, it is specialised in several fields and has a widely qualified staff, which confers a very competitive advantage to the group by qualifying us to take on projects that involve many different disciplines. This also enables us to take on projects from different areas and, what is even better, is that we can provide knowledge and quite eclectic points of views to solve diverse scientific and technological problems. It is worth highlighting that the important bit is not the multidisciplinary nature of the projects but the fact that we have many specialists from several disciplines at our disposal and therefore the possibility to collaborate with ease thanks to our common trajectory and experience. We believe that this is the key to the success of the initial approach of the GII, which had sought this multidisciplinary consolidation that has taken us many years to achieve.
(Q).- In this sense the Campus do Mar project, promoted by the three Galician Universities, the CSIC and the IEO, together with 4 Portuguese Universities, is a good example of integrated research. What is your opinion on this initiative?
(A).- We surely think that this is an initiative of great interest for the theme it covers. In this sense, and right from the beginning, we have always expressed willingness to collaborate with the Campus do Mar to help it progress towards success which it rightly deserves, in sectors that we think are basic for Galicia and in which we could be joint leaders. The GII has always felt that collaboration and integration, in the way it is planned by the Campus do Mar, will certainly lead to achievement of the most interesting goals rather than when one acts in isolation. This has led us to choosing a collaboration policy with a huge number of Galician, national and international groups.
M. Norte/Campus do Mar