Commercial robots to drive industrial applications

Companies will ultimately seek to build humanoids that can effectively replace human beings in performing multiple routine tasks.

by Andy Wilson
EDITOR
andyw@pennwell.com

Recent developments in robotics and machine vision have led to the introduction of systems that target diverse markets such as consumer electronics, entertainment, military applications, and personal assistants. Such systems have appeared in the form of automated vacuum cleaners, pet dogs, robotic fish, reconnaissance robots, and humanoid robots. Many of these systems incorporate multiple sensors to recognize, interpret, and understand human actions. After training, some autonomously perform tasks in hazardous or dangerous areas or do repetitive everyday tasks.

Catching the headlines in last year's popular press were two products aimed squarely at performing everyday tasks: the inexpensive Roomba from iRobot (Burlington, MA; www.roombavac.com) and the expensive Trilobyte from Electrolux (Stockholm, Sweeden; www.electrolux.com). Both products use navigation technology to automatically clean household floor surfaces.

Users can now spend less time cleaning and more time playing with the latest robotic entertainment sensation from Japan—Aibo, the intelligent dog. In its latest incarnation, Models ERS-210A and ERS-210 can autonomously recharge themselves and then continue their autonomous behavior. "By infusing Aibo with increased artificial intelligence, such as voice and face recognition, we are expanding its autonomous functionality," says Victor Matsuda, vice president of Entertainment Robot America, a division of Sony Electronics (Park Ridge, NJ; www.sony.com).

Those who do not like dogs may wish to consider a robotic fish. Such a fish was developed in a $1 million, four-year research project at Mitsubishi Heavy Industries (MHI; Tokyo, Japan; www.mhi.co.jp), a company better known for building ships and submarines. According to a report

by the BBC (London, England; www.bbc.co.uk), MHI has created a robotic replica of the rarely seen coelacanth, a fish once thought extinct. MHI expects the technology to be used in virtual aquariums or in autonomous deep-sea research vessels.

Such robotic developments will certainly drive future initiatives in bringing together the once disparate fields of machine vision, artificial intelligence, voice recognition and synthesis, and autonomous motor control. In doing so, companies will ultimately seek to build humanoids that can effectively replace human beings in performing multiple routine tasks.

Recognizing this trend, the US Defense Advanced Research Projects Agency (DARPA; Arlington, VA; www.darpa.mil) has started a research initiative in the field of cognitive systems by releasing a Broad Agency Announcement that seeks innovative research proposals. The DARPA Cognitive Information Processing Technology Initiative will develop cognitive systems that reason, learn, adapt, and feature self-awareness functions. DARPA's vision is to develop systems that can overcome these problems by being responsible for their own operations and be able to cope with unforeseen events. They will reason, using appropriately represented knowledge, learn from experiences, and be aware of their own behavior.

But this move by the US government is a shade late. Japanese companies that have developed and deployed numerous systems in hundreds of applications now dominate the humanoid robotics market. As for an autonomous cognitive machine, the Honda Motor Co. (Tokyo, Japan; www.world.honda.com) intelligent humanoid robot Asimo (Advanced Step in Innovative Mobility) is already capable of interpreting the postures, gestures, and voice commands of humans and moving independently in response.

Since its introduction, Asimo's capabilities have advanced significantly; it can greet approaching people, follow them, move in the direction they indicate, recognize their faces, and address them by name. Honda plans to rent the humanoid as an automated receptionist. Rather than re-invent the wheel, DARPA's money may be better spent investing in development of already successful systems.

More in Non-Factory