Making sense of sensors

Accenture is watching you.

Somehow, that statement just doesn't ring with ominous, Big Brother overtones. The consulting firm, however, is placing a big bet on remote devices and sensors that will gather information on the location, status and temperature of millions of objects in the world, or of their surrounding environments. Though some say these systems could corrode individual freedom, others believe they'll give people and businesses important information they can't get now. Accenture doesn't make the sensors; instead, it prowls the labs of other companies and tries to come up with ways to weld disparate technologies into a cohesive whole.

Accenture's Chief Scientist, Glover Ferguson, is in some ways the head prowler. For the past few decades, he's worked at the company trying to figure out what's next. Ferguson sat down with CNET News.com to discuss RFID, the general state of privacy and what your car radio is saying about you.

Q: Give us a quick run down on Accenture's lab.
Ferguson: It first got started when Accenture was still part of Arthur Andersen & Co. We were about to get into the software business in a big way, and the argument was that we can't be in the software business and not have a lab. It's just illogical. So, the first site was in Chicago. Why in God's name would you put a research lab in Chicago? We actually had a headquarters there, and the thought was that if we put the researchers anywhere else, it could be too easy to forget that they existed.

It was the first organization to break the dress code. They said you can't have researchers in suits, because no one will believe they're actually researching anything.

What sort of projects do you tackle?
Ferguson: One of our charters is to construct a five-year moving vision as to what we think is going to take hold, with the goal to create a working prototype.

In 1997, we started looking at some early RFID chips. RFID is actually a WWII era invention for identifying friendly aircraft. So, they did commercialize it, but on very, very high-end assets. We asked if you could drive it down to revolutionize supply chain. The answer was yes, but it would have to wait on standards. When EPC global got started, that started to shape up.

I think we still have a lot of discussions to get over before we get comfortable with where this is going.

Anyway, the first prototypes tend to be done by research and tend to be pretty quirky because they speak to as broad an audience as possible. The first one in RFID was a talking medicine cabinet. The first thing it does is recognize you; there's a little facial recognition thing going. It's an entirely local application, so you're not out there (on the Internet). It recognizes you and says, "Good morning, the pollen count is pretty high today, you'd better take your allergy medicine."

So, you reach in and grab a pill bottle and pull it out and that's where the RFID is. It says, "That's not yours, that's the wrong medicine," and you put that back, pull out another one and it says, "That's the right one. Now take two of those." And the mirror--instead of just being a mirror, it has a screen. All of these things are now starting to be actually discussed commercially, but in '97 it was "Oh, come on."

We showed it around and we got questions like, "So, Accenture is going into the medicine cabinet business?" No, we said, "Look at the capabilities this is demonstrating." One of our operating group said, "You know, my client sort of gets it, but it'd be better if we could particularize it for his business," His business was gas cylinders. So, the next one we built was an RFID system to track the life cycle of a gas cylinder: everything from filling to "do not fill this tank with that gas," to knowing where the inventory was.

Industrial customers have flocked to RFID, but consumers still have a lot of concerns about privacy.
Ferguson: Consumers are willing to sacrifice some of their privacy, with two caveats. One, they get something for it. Two, they understand what, exactly, they've given up. If you take the data and do something else with it, you violated one of the rules. If you take their data and don't give them anything for it, but just use it to enhance your own profitability, they don't care for that either. But after that, if you don't do those two things, it's not such an issue. I think we still have a lot of discussions to get over before we get comfortable with where this is going, and that's what you do in a society to get the vote on these things. California will probably start with laws that are too far to one side, but eventually we'll end up with something we're all comfy with.

I used to be teased by my British colleagues about how sloppy Americans are with their personal data on Web sites. "We in the U.K. value our privacy. You Americans give it up all over the place. We value our privacy so much we don't even have pictures on our driver's license."

I took this for years, then one day I said, "Wait. London has more CCD cameras per square inch than any other city in the galaxy. How dare you talk to me about privacy!" They said, "That's different. That's security." Where do you draw the line? Americans will give up data if

they think they are going to get the convenience; sometimes they do, sometimes they don't. In those days of the Irish troubles, (people in the U.K.) craved a higher sense of security and they permitted all these cameras to be put up.

There does seem to be a lot of passion when RFID comes up.
Ferguson: But it's funny. A lot of these discussions get very holy and people pace back and forth and give their speeches, but they keep stepping over this dead horse in the middle of the room. It turns out that most of us are living some aspect of our life slightly outside the law. Maybe I drive a little bit over the speed limit. There are people who take aggressive positions on their taxes. There are lots of people who do something that by the letter of the law is not legal, and frankly, they don't want to get caught. We can turn this into a very, very holy speech on privacy instead of discussing whether maybe we shouldn't speed or maybe the speed limits should be higher.

And in some cases it's just people not understanding, I hear a lot of concerns about RFID where people worry about being scanned and someone getting all their personal details. No, they will get a 91-bit number. Now all they have to do is hack into 17 different databases to find out anything about you.

People are afraid someone will drive by their house and scan everything in it. No, if they had a scanner powerful enough to do that, your fear should be being cooked in your own home.

When we get enough sensors planted out there in the world, what are they going to do? A lot of people like Dust Networks and Crossbow are employing them to reduce energy consumption in buildings.
Ferguson: We've described it as creating a virtual double. If I get enough feeds from enough different directions, I can start to construct virtual versions of every physical and real object on the planet, and with those I will actually have more information and control than I have if I confront the object itself.

So, for example, if you confront a printer, you can figure out right away whether it has paper or toner and whether or not it's working. But if you stayed at your desk, and there was a virtual double, you could have had those two facts, plus you could see that some idiot sent a 5,000 slide PowerPoint to that printer.

What other novel applications for this sort of thing do you see out there?
Ferguson: One shopping mall has a billboard that's listening to the leakage off FM antennas of the cars driving by. From that it can determine what people are tuned to. Now, there is a privacy issue, but I don't know who the hell they are. I just know that at this moment in time, people who can see this billboard are listening to Montovani or that people who can see this billboard are listening to rap.

Where do you draw the line? Americans will give up data if they think they are going to get the convenience.

The billboard then changes its message based on who's receiving the message. Why would I try to sell ocean cruises to kids, or skateboards to older people? The benefit to the driver is they actually see something that's applicable to them. The benefit to the advertiser is that they get their message to the right people. And the guys selling the space can charge more because it is a particularized message to a particularized segment.

Did you work on that project?
Ferguson: No, we didn't. The frustrating thing about a lot of the stuff we work on is that we can't get clearances to talk about it.

How quickly will the sensor world evolve?
Ferguson: I think with this particular vision, we got out in front of ourselves a little bit. We're still working out some of the physical issues with RFID and reading through liquids and metals and things like that. Aside from that, there is a missing underlying infrastructure. We can put devices everywhere, but then you have to be able to provision them, know when they're sick, know how to fix them. You'll see lots of it within the next five years, but things will only really take off, at least in their full flower, once you can assume the infrastructure is there.

We have a vineyard up north and the idea is to optimize the output. If we can report on the microclimate of a vineyard, we can start to optimize how each portion of the vineyard is utilized. We can measure sunlight, we can measure leaf moisture, soil moisture, temperature. After the sensors started gathering data, I think they (the people running the project) told me that something like two thirds of the data coming in was these things saying, "I'm out of battery. I'm broken. I've lost contact."

I recall being at an Accenture event five years ago and one of the consultants showed me an experimental videoconferencing system, and the consultant said that one of the problems they had is that people in tests said they didn't like getting incoming video calls.
Ferguson: We have a much more casual system here. We have a video tunnel set up (between Chicago and Palo Alto, Calif.), so researchers can see each other walking by like they would in a hall. They'll stop and just have a face-to-face conversation, so it's a different paradigm. And of course we built a prototype for that, but when you saw things, the cost of doing that would have been crazy.