CitationDownload as .RIS
Emerald Group Publishing Limited
Copyright © 2004, Emerald Group Publishing Limited
The future is now?
The future is now?
Keywords: Research, Robotics
Giovanni Muscatois based at the Dipartimento di Ingegneria Elettrica Elettronica e dei Sistemi, Università degli Studi di Catania, Catania, Italy.
Around the turn of the 17th century the Italian philosopher Gianbattista Vico developed the idea of the cyclical nature of history. According to Vico, history repeats itself in such a way that, although with increasing complexity and ever changing features, events are driven by the same underlying principles and eventually produce the same results.
This idea has undoubtedly some merits: it can interpret remarkably well past historic events in robotics and with a bit of visionary input it could be perhaps a useful tool to develop hypotheses for future developments.
We assist cyclically to periods of incredible prophecies of a futuristic society with many robots in everyday life and periods where the sceptics prevail. Will the next future world be dominated by the presence of robots or not?
If we look at the past we can better understand the future. A dishwasher machine or a washing- machine or a coffee-machine are simple examples of robots that are commonly used in everyday life. They do not walk or smile, do not use hands, but perform their job. I want to extend some personal opinions concerning the actual world of robotic research that I (roughly, and with many intersections and overlapping) classify into the following four areas.
Academic basic research. New problems and solutions are proposed, but very often formalised in such a way that no real system will satisfy the hypothesis. Sometimes I have seen researchers working in robotics that have never seen a real robot, but only a set of equations. Basic research is fundamental, the sky is blue, but we walk on the earth.
Futuristic based research. This is followed by professors or scientists who like to give prophecies on the future of the world and that promise that in a few years incredible robotic systems will be available in our everyday lives to solve most of our problems. Nobody asks which actuators and computing systems will be adopted (because these will be available in a few days...), but many journalists like to write what they say. Only simulations of the systems are shown or simple prototypes, with the sentence: “in the next few years this will be able to recognize our speech, and to autonomously perform its job”, or “it is actually like a 1 year old children, he needs to learn and then we will see what he can do”. In the meantime the years go on.
In one way, the excess of fantasy is good, because it creates new ideas, but in the other the disappointment created to the public as a consequence that actual systems are much simpler, is dangerous for the robotic community. Fifty years of science fiction have produced in the society expectations for robots that are far from the real products, and this leads to increased scepticism. We should reverse this process.
Industrial product based research. “The system must be ready yesterday” is the basic rule. The solutions found are often original, however these are not easy to be generalised. It works, but nobody knows exactly or can prove why it works, from a rigorous point of view it is not reproducible, so it is not science.
Application specific research. This comprises the development of systems in some domain where a direct and ready solution is not available in the market and new products and applications are created. Very often the developed system does not solve the problem in a satisfactory way or is too expensive for introduction into the market.
In my robotic experience I have faced several challenging applications, such as building robots for picking fruits, for climbing walls, for inspecting microareas, for going inside volcano craters, just to name a few. However, even if I appreciate basic research and emerging technologies, I do not approve of those who promise to build in a few years something artificial that will reproduce human life, or that will be capable of thinking autonomously. In my opinion we are too far from that.
It is true that many new technologies have been presented, but these are very specific and deep knowledge and a long time are required to understand their real possibilities.
For example, the computational power that is currently available is really incredible and is under exploited in most robotic applications. However, from what we have seen in the past, it is not simple to understand if a completely new methodology for developing software or more powerful hardware is needed. The number of processing units and the connectionism of a biological brain is in reality very far from actual technology, and the understanding of the biological mechanism of thinking is at an early stage. Moreover will we be ready to accept a robot that needs 1 year learning to start walking, 5 years to start writing etc.? Biological inspiration does not always provide the ideal solution. The example of airplanes that are faster than birds and do not flap their wings, is always appropriate.
A real explosion of new robotic applications will happen probably when new actuators and power supplies will be available. Smart materials like IPMC and Nitinol, or power sources such as fuel cells were announced as the solution to everything, but a lot of research is still needed.
Many new sectors are just around the corner; in the near future the best chance will be to adopt existing technologies to solve problems in different fields.
The market for robots is increasing each year and there are good opportunities for new applications. This means that very simple actions will be performed by robots for two main reasons: the first one is the decrease of cost that allows a wide adoption of new systems; the second is due to the fact that all the new generations are now more familiar with the new technologies. For example, the wide adoption of robotic systems in the medical sector is also due to the fact that several doctors have only recently opened their minds to the adoption of robots.
On this point a crucial role will be played by the increase in modularity that will allow the systems to be easily reconfigured. New standards are needed; those who affirm these standards will win. If we increase the levels of modularity and introduce new standards in robotics, new markets and applications will come automatically. We do not need to think about totally different systems, but how to introduce cheaper and more reliable systems in everyday life. Service robots are typical examples of a class of products announced several years back with incredible perspectives. These expectations have not been reached, but were useful to establish a series of new applications that are slowly emerging.
During the last CLAWAR conference (at which many papers in this issue were presented), I was asked by several journalists about what new revolutionary products were announced by the attendees. Probably we have not seen anything really new, but what we have seen is that many new robots are ready to go out from research laboratories to enter everyday life. Little by little, leaving the door open, in any case...