Couple of weeks, ago, I had my first symposium on line ever. This was organized by eSymposium and gathered mainly Academics but also some people from VCs and financial community.
Industrial robots
Robots have long been used to manufacture cars - for the moment, one robot might be dedicated to specific tasks and much is expected on modularity e.g. the ability to reprogram and reconfigure a robot for new tasks. This would allow companies to better leverage their investments and adapt quickly their workforce.
Much has been said on future "service robots" that will capitalize on human intervention, helping learning how to behave, manufature, etc. Today, more than 80%of all robots are used for cars, but many of them might be used for other sectors in the future.
Honda, Toyota, etc. all spend much on robots R&D - emphasis is given on humanoids. One ambition is to replace workforce for all tasks. Today, robots are used only for painting cars and handling non precise tasks.
Residential robots
Much emphasis on vacuum cleaner robots. But many researchers invest time and energy on location and navigation. This require huge efforts and some might ask "is this worth ?"...
A guy from Sony / AIbo was also there - he emphasized much on the ludic side of their pet. Sony does not target a more limited robot (navigation is a big chunk of the pet's cost) and thinks that economies of scale should help marketing Aibo to a broader audience in the short term.
Artificial intelligence
Many question remain on how to integrate all information in one box - it seems that NASA has launched a program on that (human robot interaction), but participants emphasized the need for "ambiant intelligence" in the short term. Current models do not have the necessary processing power to compute all information and should rely on the environment in the short term to appear as intelligent as possible.
However, everything is still at the prototype stage and one might be amazed by the limitation of current robots. Worth mentioning that most of the robots shown today are controlled by external servers and are rarely completely independant from human...
Some think that studying human / animal interactions should help build up a language for programming human / robots interfaces. Interesting...
Worth / to keep in mind: Philips Research releases a platform called iCat that intends helping researchers understand and enhance human / robots interfaces. Nothing new, except that this is cheaper, nice, simple - but all the big guys make their platform "free" to researchers. Nothing is free in our brave new world: they all would like to have their platform adopted as a standard !