Roaming With...Wireless: It's Not Just for People
By Herbert Schorr, Executive Director, USC Information Sciences Institute
When we say “wireless” most people think about their wireless phones – about teens texting, or Mom and Dad calling each other to coordinate a soccer game pick-up. But as important as those calls might be, wireless is about more than people talking to each other. Wireless technology is also connecting machines that, by virtue of the evolving capabilities of integrated networks, can provide enormous advances in a number of areas.
Here at the University of Southern California Information Sciences Institute (ISI), which has pioneered such key wired systems as the Domain name system and open source Grid computing, we are pursuing basic research on wireless, which encompasses a vision of a wireless that doesn’t link individual human users, but rather swarms or committees of machines.
The earliest applications for many technologies have long been military and scientific, for example seeding a wide area with modules designed to listen for sounds of vehicles, which could listen for weeks, months or even years at a time, to track vehicles heard, deduce their location and movement, and notify operators.
But opportunities for commercialization of the ISI’s wireless research are now emerging – with the first examples likely in the field of oil exploration and extraction.
ISI computer scientist John Heidemann, with numerous collaborators, has for years been exploring the idea of ‘sensornets’ – a collection of autonomous units set out in an array to continually monitor conditions in off-the-grid environments.
The challenges, hardware and software, of creating a useful array able to do this are formidable. Having individual units individually ‘phone home’ on a continuing basis is just about impossible for large-scale deployments. Battery life is a major constraint.
Instead, the vision was to have the individual units waking intermittently, whispering quietly among themselves, and relaying detailed messages as needed. According to Dr. Wei Ye, a collaborator with Dr. Heidemann, “the challenge is to keep energy consumption low, but still efficiently and continuously extract data when a sensor detects events. For many events you want sensors to cross check readings for validation.”
By Herbert Schorr
Executive Director
USC Information Sciences Institute
