IoT Increases Efficiency

  • Comments posted to this topic are about the item IoT Increases Efficiency

  • There are some unexpected consequences for fitness apps http://www.bbc.co.uk/news/technology-42853072

  • I saw that. A very unexpected consequence, and certainly a reason why we need better data understanding and controls.

  • Unfortunately the universe of "unexpected consequences" is virtually infinite.

    ...

    -- FORTRAN manual for Xerox Computers --

  • I was approached a few years back about this problem for a restaurant.  They wanted to outfit everything with sensors everywhere and a new point-of-sale(pos) that also integrated inventory control systems.  Their PoS would display items as available when they really weren't.  We did integrate a punch of RaspberryPi devices outfitted with all the sensors they wanted, not just needed.  Everything was streamlined and even added a few thermal printers to tag items as they came in to maintain the FiFo and reduce the back stock they had.  Instead of centralized ordering kiosks, every employee got an iPod because trendy.  They were able to recoup costs by the reduced food waste, and schedule employees better too.

    Sadly before the project could move into its next phase I moved far enough away to not be an active member.  They were going to add sensors to the kegs to weigh them when they came in to measure just how much was left and display it above the bar.

  • In the UK the electricity meters get read once a quarter.  Smart metering (if the public accept them) enables continuous monitoring and potentially down to specific devices.
    In theory it should help power planning.  In practice I'm skeptical.  I've visited a few power stations and they are pretty good at predicting demand.  People are creatures of habit.
    Apparently the next generation of cars will generate 2TB of data a day.  God knows what that will be.  Again there are far less intrusive ways for potential users of that data to get what they need.  Perhaps a different approach will make that 2TB a day irrelevant.  For example http://www.streetbump.org/about

  • David.Poole - Wednesday, January 31, 2018 12:44 AM

    In the UK the electricity meters get read once a quarter.  Smart metering (if the public accept them) enables continuous monitoring and potentially down to specific devices.
    In theory it should help power planning.  In practice I'm skeptical.  I've visited a few power stations and they are pretty good at predicting demand.  People are creatures of habit.
    Apparently the next generation of cars will generate 2TB of data a day.  God knows what that will be.  Again there are far less intrusive ways for potential users of that data to get what they need.  Perhaps a different approach will make that 2TB a day irrelevant.  For example http://www.streetbump.org/about

    Organizations, governments and companies are becoming drunk on data. Magic cure-all.

    [A while back a Ford executive got a lot of backlash for quipping 'we know when you're speeding'. But the implications buried in that statement are frightening.]

    ...

    -- FORTRAN manual for Xerox Computers --

  • Collecting all this data on your car worries me as we already get errors and misinterpretations with current sensor systems. Recently I got a small leak in a vacuum pipe - this generated an error which then cascaded down through the system to be finally reported as an automatic gearbox problem. A friend's car is forever telling him that he has an OSF puncture when he hasn't!

    Also there is a "data for data's sake" mentality. I spoke with a supposed IoT guru a couple of years ago and he was keen to collect ever more data (even increasing the sampling rate). My background makes me look for "relevant" data potentially disregarding a large percentage as "irrelevant noise". There was no way we would ever agree!

  • More data means more irrelevant data, but it also means we need more complex analysis and not the simple one error is an issue.

    A guy from CERN talked about this a few years ago. They get 1PB/s in some HARON experiemnts and sensors need ot decide what data to keep and drop because computers can't handle that rate. However, they also have to learn to account for some bad data. Part of the IoT issue now, and even OLTP issues, are that too many of us assume all data is good and don't necessarily account for errors.

Viewing 9 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic. Login to reply