More on The Internet of Things: It may appeal to you if you’re a visionary or a systems company
Several (all right, all) of the comments on the post, “HP announces sensors for its “Central Nervous System for the Earth” (CeNSE),” raised the question of exactly what is the purpose of having a trillion (literally) sensors out there monitoring the world? I had attempted to gloss over this question by linking to a Wikipedia definition of “The Internet of Things,” but I think it’s worthwhile going into a bit more depth on the concept, if only because of the implications it has for future sensor applications.
The Internet of Things is the networked interconnection of everyday objects – soda cans, shoes, refrigerators, you name it – mostly through RFID tags and IP addresses. Here’s a good example of one implementation of it: Ford’s Tool Link system (a $1220 option) has sensors built into vehicles like the Ford Transit Connect so that when the driver presses a button the van displays an inventory of all tools onboard. A similar system for the home would show you an inventory of all clothes that were supposed to be in your suitcase, or things in your briefcase.
So HP’s CeNSE approach is a step further along in sophistication. Stuff can communicate information about itself, such as its temperature, or its motion (vibration) over time. HP isn’t the only company working on it – IBM, for example, calls its program ASmarterPlanet. The goal for the Internet of Things is that all Stuff is instrumented, interconnected, and intelligent. The Bay Bridge linking San Francisco and Oakland was closed a couple of weeks ago after a piece of it fell onto the roadway. Presumably this wouldn’t have happened in a world shaped by the Internet of Things: Thousands of accelerometers on the bridge would have registered the vibratory signature of impending failure.
Are we there yet, or even close? No. Is it likely we’ll achieve this new interconnectedness? Companies like HP and IBM, which deal in hardware and software that scale massively, are putting a whole lot of effort into seeing that we do.
Policebox commented:
Meredith Poor commented:
One approach for setting up an efficient robotic lawn mowing system is to divide the yard into 'tiles', although tiles in this case overlap, and are designed to enclose the complex paths that are followed on a curved sidewalk, for example. Any particular tile has a set of reference waypoints, whether RF, IR, visual, or sensor. If the robot is cutting a straight swath the tile runs the length of the swath, and the reference points are accessable from every point on that swath. When cutting around a curved flower bed, a number of tiles may overlap, and any particular cut may be no more than a foot. Each shrub, tree, mailbox post, and lamppost may have an RFID for signaling proximity to the robot.
This is not simply a matter of 'taking inventory', it creates a foundation for acting on those objects through service robots, cleaning equipment, or disposition technologies. Which is easier, trying to figure out whether a can is steel or aluminum, or simply reading it's ID?
