AQUA, an amphibious, otter-like robot, is small and nimble, with flippers rather than propellers, designed for intricate data collection from shipwrecks and reefs.

The robot, a joint project of 91ŃÇÉ«, McGill and Dalhousie universities, can now be controlled wirelessly using a waterproof tablet built at 91ŃÇÉ«. While underwater, divers can program the tablet to display tags onscreen, similar to bar codes read by smartphones. The robot’s on-board camera then scans these two-dimensional tags to receive and carry out commands.

Cutting the cord on underwater robots has been a long-standing challenge for scientists; water interferes with radio signals, hindering traditional wireless communication via modem. Tethered communication is cumbersome and can create safety issues for divers.

“Having a robot tethered to a vehicle above water creates a scenario where communication between the diver, robot, and surface operator becomes quite complicated,” says (right), professor in 91ŃÇɫ’s Department of Computer Science & Engineering in the Faculty of Science & Engineering and co-author of the forthcoming paper, "Swimming with Robots: Human Robot Communication at Depth".

“Investigating a shipwreck, for example, is a very delicate operation and the diver and robot need to be able to react quickly to changes in the environment. An error or a lag in communication could be dangerous,” Jenkin says.

Realizing there was no device on the market that fit the bill, Jenkin and his team at 91ŃÇɫ’s Centre for Vision Research, including the paper’s lead author, master in computer science student , set to work constructing a prototype. The resulting device, fittingly dubbed AQUATablet, is watertight to a depth of 60 feet. Aluminum housing with a clear acrylic cover protects the tablet computer, which can be controlled by a diver using toggle-switches and on-screen prompts.

“A diver at 60 feet can actually teleoperate AQUA 30 to 40 feet deeper. Needless to say this is much easier on the diver, physically and much safer,” Jenkin says.

The tablet also allows divers to command the robot much as if they were using a video game joystick; turn the tablet right and AQUA turns right, too. In this mode, the robot is connected to the tablet by a slim length of optical cable, circumventing many of the issues of a robot-to-surface tether. The optical cable also allows AQUA to provide video feedback from its camera to the operator. In a totally wireless mode, the robot acknowledges prompts by flashing its on-board light. Its cameras can be used to build 3-D models of the environment which can then be used to guide the robot to particular tasks.

“This is a huge improvement on [a robot] having to travel to the surface to communicate with its operators,” Jenkin says.

In the past, divers have used laminated flashcards to visually communicate with robots while underwater. However, these limit the diver to a pre-set sequence of commands.

“It’s impossible to anticipate everything you’re going to want the robot to do once you get underwater. We wanted to develop a system where we could create commands on the fly, in response to the environment,” he says.

Jenkin and Verzijlenberg’s paper will be presented at the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) in Taiwan.

Jenkin and Verzijlenberg are two of the researchers based in 91ŃÇɫ’s new state-of-the-art Sherman Health Science Research Centre, which officially opened on Sept. 14. Jenkin leads the Canadian Centre for Field Robotics, which is based on the building’s main level. The centre is supported by a grant from the . The AQUA project is funded in part by the . 91ŃÇÉ«'s Centre for Vision Research is part of the Faculty of Health.

Republished courtesy of YFile– 91ŃÇɫ’s daily e-bulletin

The post Dream of exploring shipwrecks? Check out Professor Michael Jenkin's wireless controller and underwater robot appeared first on Research & Innovation.

While current search-engine technology uses algorithms that rely on keywords to uncover pertinent information, the research of Huang and his colleagues uses algorithms with a capacity for natural-language processing and contextual understanding that can uncover more complex information. Rather than looking for instances of words, these elaborate algorithms can make sense of the words. The applications of the research are vast.

Right: Jimmy Huang

They can analyze, for instance, the text of blog postings and the feedback of its respondents and then make humanlike, intuitional assessments of the information. “The analysis could indicate changes in public mood on certain issues, or rising or diminishing support for a certain political leader,” says Huang.

His technology research seeks to improve the health-care system by reducing costs and improving services. Huang advocates the digitizing of all health-care records – texts, charts, X-rays and other image files – and making them securely accessible on the Internet.

How will this improve health care? “One of the reasons that health care is so expensive is due to the incentive for doctors and specialists to provide unnecessary or duplicate services, and to the inefficiencies of a system that creates multiple, proprietary medical records,” says Huang. He cites the example of a general practitioner drawing blood for a test, then referring a patient to a specialist who might perform the same procedure again due to the inaccessibility of the original GP’s files.

“If there were one comprehensive set of records for the patient, centrally and securely available on the Internet, practitioners wouldn’t have to resort to phone calls and courier services to exchange information,” says Huang. Improved access to information could reduce services and associated costs while maintaining the same level of care. “It would also allow for a more global perspective on a patient’s health as various practitioners wouldn’t be limited in the scope of information available to them and, with the assistance of algorithms, could uncover heretofore unseen conditions.”

Huang offers another example of the benefit of centralized resources by citing his own experience with a leg injury. His doctor referred him to a specialist near his home. However, the specialist couldn’t fit him into his schedule for two months. By then, the injury was repaired on its own, for better or worse. A centralized information system could analyze a patient’s surroundings in a more subtle way, not simply in terms of where the GP’s office is situated, but where the patient lives or works. An advanced algorithm could locate an available specialist in another area who might be available sooner.

The sphere of application is, likewise, immense: from opinion mining and sentiment analysis to context-aware computing and social networking or matchmaking.

The joint 2010 International Conferences on Active Media Technology and Brain Informatics will be on the Keele campus from Aug. 28 to 30, and the 2010 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology will take place at 91ŃÇÉ« from Aug. 31 to Sept. 3. The 19th ACM International Conference on Information and Knowledge Management (CIKM 2010) will be held at the Fairmont Royal 91ŃÇÉ« from Oct. 26 to 30.

“Bringing these venerable, annual conferences to 91ŃÇÉ« for the first time represents quite a coup for the University. Among the list of sponsors are some of the major companies involved in the information technology industry and the competition for papers among prospective participants was extremely competitive,” says Huang. “I’m very grateful for the support of the Offices of the Vice-President Academic & Provost and the Vice-President Research & Innovation, and the Dean’s Office of the Faculty of Liberal Arts & Professional Studies.”

Republished courtesy of YFile– 91ŃÇɫ’s daily e-bulletin.

The post Professor Jimmy Huang to host three IT conferences; topics include artifical intelligence appeared first on Research & Innovation.