Sensors Expo Fall 2003: Sense the world through online eyes
Anaheim, CA—The interactive and increasingly pervasive nature of the Internet, sensing devices, and connecting communications were some of the issues covered by Lucian (Luke) Hughes and Gary Boone, both of Accenture Technology Labs, (Palo Alto, CA), when they delivered a Sept. 23 keynote, “The Sensated World Brought Online,” during Sensors Expo Fall 2003.
Anaheim, CA— It won’t be long before the Internet blends so seamlessly with our daily lives that we’ll browse reality in a digital way. Talking about “going online” will seem as odd as talking about “getting power” would be now for abundant electrical devices. No one thinks or talks about it—it just happens.
Further, business decision-makers don’t care about readings of individual sensors. Instead, they want information to help them make smarter business decisions—real-time yield maximization. The interactive and increasingly pervasive nature of the Internet, sensing devices, and connecting communications can help make that possible. These were some of the views presented by Lucian (Luke) Hughes and Gary Boone, both of Accenture Technology Labs , (Palo Alto, CA), who delivered a Sept. 23 keynote, “The Sensated World Brought Online,” during Sensors Expo Fall 2003 .
Hughes and Boone, among 60,000 people at Accenture, help clients see how technologies impact research, development, and business development, primarily in the 0-5 year timeframe. They say that mobile phones, PDAs, and cars and trucks (along with cameras, Internet access, and global positioning technologies) already help people interact with reality in a digital way.
They add that lower prices lead to pervasiveness. Radio frequency identification (RFID) technology, once about $50 per tag and used primarily on railroad cars, is soon to be everywhere, Hughes reports, citing a recent sale of 500 million tags at 8-13 cents each. “Wireless dust,” a sensor with power source and transmitter, once $100-$200 per speck, might get to $1, allowing a cloud of them to be dropped into various locations, such as a battlefield.
Turning data into information for good business has recent examples, Hughes says. One is an increased effort by airlines to monitor and treat limited assets—seats on a plane—as demand-based commodities with appropriately changing prices. Agricultural equipment uses satellite data to detect moisture levels and infestations, and then automatically applies water and pesticides exactly where needed. A Mexico City cement company reduced late delivery average from three hours to three minutes by circulating full cement trucks in heavy traffic around town, selling and dispatching trucks to sites as needed, based on the expiration time of the assets and their locations. That’s more efficient than loading and sending trucks only after an order comes in.
In one of the most dramatic instances of real-time (or real-time enough) data mining is Wal-Mart, explains Boone, which has a data center with 100 Terabytes of capacity. Cash registers take weekly inventory from 3,400 stores, compare demand to stock and trends, and make profitable decisions with help from appropriate algorithms and an 1,800-person IT department. The demand-pull system makes about 90% of the decisions related to about 700-million store/item combinations, he says. In 1999, for instance, purchasing data revealed a flu outbreak, and Wal-Mart stocked shelves appropriately in advance of its spread, creating customer good will (and good business), from a bad situation. Hughes admits that the concept isn’t new but the tools are better. For instance, people in WWII fighter command centers used information called in from spotters to move model planes around a large board to make better deployment decisions more quickly.
Where’s it going? Sensor enterprise systems will help aggregate data into information, Hughes adds, for a “reality online big picture.” Boone did a demo where sensors in a forest could tell where fires start or when wood should be harvested based on grade and market conditions, or, when an individual premium wine grapevine (top grade wine grapes sell for 50-100 times the price of raisins) need specific attention. He also expects maps and other web services to merge, offering much more useful zoom-in maps of restaurants with pop-up signs and click-in menus. Data can also be mapped in aggregate to show regional buying potential based on employee optimism, a better indicator of spending trends than some other metrics, he suggests. “The sensors themselves don’t matter; it’s how you visualize the data and act on it,” Hughes says.
A personal reality assistant is in prototype testing now, where the user can input info into a carry-around computer, and then retrieve names or facts upon request. Such tools will become natural extensions of the 350,000 e-mails sent daily in the U.S. alone.
Hughes notes, and audience discussion afterward suggests, that tolerance for privacy-invading technologies will be tested over time. People may continue to “pay” a little personal information to trusted third parties in return for receiving valuable services. For instance, at one security-based location, those with an option enter in three seconds with a fingerprint reading, as opposed to a 30-second alternative security procedure. Areas of likely rapid growth for transparent networks include security (40% of $2-3 billion being spent on a U.S. Security sensor network will be for detection systems) and Web-phone enabled services (qualified insurance representatives may send accident photos to a central location of insurance experts). Further, when RFID tags are in all items of value, theft might get to be nearly impossible. Society will continue to decide, Hughes says, if the benefits are worth giving up specific technology-enabled information.
Control Engineering Daily News Desk
Mark T. Hoske, editor-in-chief