Applications of IOT Technology in the Supply Chain

okay go ahead and get started welcome to the first SPL seminar for 2018 and today's topic is the Internet of Things and big data analytics for supply chain and logistics our next seminar will be February 20th by professor Sebastian hakuta and the topic will be applications of machine learning in the supply chain and that seminar will take place at the same time same place on the 20th our speaker today we're very fortunate to have Professor Russell Clark rest us a little bit of everything everywhere Russell's very involved with our Savannah campus our Atlanta campus but he's big into IOT research into supply chain applications with the Smart City applications you may met dr. Clark visit and expound upon it please join me in welcoming all right welcome okay yes my name is Russ Clark I am both a research faculty in computer science I'm also a Norfolk Tech alum from the grad school many years ago I come on my background is computer science on the networking guy I teach computer science networking courses mobile app development and a number of other things over the years in particular in most recent years as we've done more work in mobile that and and you know combining the mobile and the networking into this thing called the Internet of Things obviously it's a natural extension of that work in the natural application of been doing over the years and so this presentation just set the context I'm I'm a computer scientist I'm a networking guy coming to talk about what I think are some of the cool things we can do and apply this work to supply chain I thought the supply chain expert all right you guys are the supply chain and logistics experts in the room and I'm here basically to share some thoughts and share some experience and and hopefully your minds thinking about I wonder if we can use some of that in the work that we do we can apply that and some of the things that you do so as tip said I work in a lot of things related to the education space both with professional education with our campuses around the world at Georgia Tech Lorraine and France we're teaching the mobile apps semester and also starting a new program in Panama just to set the stage and something you know most of your from Georgia Tech you know a lot of the context but you don't realize all the things that are going on in Texas to sort of bring all this together right bring and so we do have a research center in Georgia Tech around Internet of Things called today of the Center for development and application of Internet of Things technologies this is a an industry consortium and industry groups that comes together on campus several times a year and we're actively soliciting student participation in that so definitely would like to talk to you more about that if you're interested Georgia Tech also is the one of the big data research centers it's called the South big data hub this is at the Science foundation-funded Research Center and Georgia Tech along with UNC are the anchors for the the southeast and so a lot of activity around big data obviously as well and as many of you know I don't have to tell you guys this Georgia Tech also has a great deal of expertise and experiences and supply chains again that's you know I'm speaking to the choir here on this topic but the other things that I know that supply chain worked as ever goes around the world the Georgia Tech and one of the key areas of interest that I've been involved in most recently is the logistics center in Panama where we have a full-time staff that focuses on very data oriented research oriented tasks and driven projects and so if you haven't seen that I encourage you to check it out and this year was starting a new study abroad program we're going to be bringing students undergrad students from Georgia Tech to study in Panama over the summer we're starting with these two courses that are really a good fit for computer science and industrial engineering students just so we can get started with somewhere between these students and certainly if you're an undergrad and thinking about or if you have friends that might be interested I encourage you to check that out all right so on to the good stuff that's the background that's the stage setter what is this Internet of Things so we know about things and we have the internet so well we start by saying it's a buzz word it's a marketing term right it's it's you know the one we use today that we used to talk about the space but it's not an especially new space in the sense that we've been doing a lot of this stuff for years you know some of these projects that we'll talk about that our IT related have roots going back 10 or 15 years and so in some sense it's you know this it's a new name from sort of person work we've been doing in sensors and you know RFID fits in this space and all these other technologies movement we've been talking about for years but but this is emphasis on that everything is being connected and so what makes it different today right but what are some things that you know why why why you're talking about this today well this this opportunity right this this space of the business this space as a researcher has has blown up in the last 70 basically the opportunity is much bigger where now we're talking on the order of 30 to 50 billion connected devices we long ago passed the point where there were more things connected to the internet and there are people on the earth right 17 billion people and now we're pushing 50 billion devices and the the money means meant of money being planned to spend of the space is huge and so that's why that's one of the reasons we're here but that's not enough the other thing is from a technology point of view this is this is you know the timing is now that which is the right time to actually be thinking and working in this space okay things are feasible today that weren't feasible five years ago in particular and we'll show some these tones I'm calling billion of research scientists I work with here has brought some toys will for show-and-tell they promote we can basically now we can we can talk about building you know deploying devices that are are fairly small but and have fairly decent radio connectivity but most importantly from an enable is going to you have really known the battery life and the sort of convergence of all those things has really changed what's possible and you know what used to be a research paper and say we could put sensors on everything or we could you know automate and you know instrument everything in this room or every package that's being shipped now that's actually feasible we're getting to the point where the technology is maturing to where we could do it so how many agree when we remember that it was only 10 years ago right what was the last code you had before that's more for the razor yes blackberry right you didn't want to call the smartphone I think that's awesome yeah before the iPhone and Android we had BlackBerry's but they weren't consumer devices we didn't all have right and it's only been 10 years since this has change you know as happen and this miniaturization the these devices now have on them a half a dozen radios and dozens of scissors right from gyros to temperature sensors to everything out and that has forced this miniaturization become automation the the focus on battery life as a precious resource right that we need to enable this I put a question mark there next information security yeah that's on purpose we have a long way to go and we certainly continue to open create challenges there all right so continuing this technology why now so low powered Wireless is one of the big enabler that is making all of our interesting ideas from supply chain and logistics projects now more feasible longer ranges and RFID and by you know models lower power consumption so years worth of battery life potentially not days or weeks and lower data rates now there's a there's a you know there's a trade-off here of these threes and three parameters right you can have a longer battery life or you can have you know faster data rates in general but what we're seeing is and especially in the supply chain application space it's that these these applications these these ideas these these uses of sensors in this technology don't require gigabit speeds right we're not going to be using the sensor to watch youtube videos on we do on our phones right did you boys cause we're going to be using this for relatively low data rate message applications like the temperature in this container has now passed the threshold or I sense that I'm near of a checkpoint afraid or or a customs check off or something like that right so and what about messages that are sent you know think I'm the urban text message and maybe only a few times a day so and when we focus on that you space if you know that application it really opens up the opportunity for what we can do and so there are several competing standards today of the space the ultra narrowband weightless and then Laura Wynn which is the one that we've been putting some work into and the kids that they'll brought to show us to play with today is is from the lorwyn technology we'll talk some more about that what we're doing on campus with that so Laura is the physical layer technology it's raining gaining traction today as an alternative to cellular and one of the places that's being deployed in this country at scale is that Comcast is using it for their machine to commercial offering which they in certain cities as trials this year and so basically they're putting this they're investing in this from a point of view of installing this the Florrick gateways the base stations in their network infrastructure right think about Comcast where does where does Comcast have invested infrastructures where do they have the footprint and in the in the game of networking it's all about footprint and where do you where do you have real estate but you might be able to add some technology where does Comcast have every home that has cable or intern and/or internet from Comcast right so imagine that set-top box or that router or whatever it is that is the Box you pay ten bucks a month to Comcast for just for the box whatever that box is imagine every one of those have one of these Laura Gateway to base stations right now we've got basically much of the country blanketed in this technology might get in this coverage so that's a big deal that a company like com and many others are looking at this opportunity so basically the way that architecture works when we talk about Laura as a protocol and the standards around Laura what we're really talking about is the interface between the end nodes a little diploma license right whatever you plug your sensor into and then the base station something like this or built into your set-top box and that is where the Laura protocols are focused and then we're talking about that back home right so wherever you want your to gather the data to to put the analytics to do the smarts and and reaching into your you know collecting the data for your applications so when we talk about Laura it's this you know three components of the architecture this is the key enabling part out here as these end notes with the assumption that we can put these gateways anywhere we have you know uplink access to the Internet and you don't have to have one everywhere so one of the applications coming back to the supply chain and logistics is this use of Laura winter access tracking in the case of so think about deploying a technology in this and in either in two ways in there a dense mode where you have lots of base stations or a sparse mode where you're really just worried about getting coverage with as few base stations as possible right and so when we talk about a dense gateway application so sparse well you know what's what's the advantage of sparse is cheaper right we can you know run as much infrastructure we can get maximum coverage for Vermaelen's yes the advantage of the dense is you're going to have more capacity on the wireless side but from an asset tracking point of view the real advantage of the dense mode is think about it as ah now I have a bunch of reference points that I know where they are and I can use those to determine my location so in general if I've got you know three four reference points three between you know three or four reference points that I can identify and use that we do this with Wi-Fi and we do this with basically just have GPS works if you want to see as many satellites as we can if I could deploy these lower base stations in a dense enough boat that my little sensor can see multiple of them now I can use that for precise indoor location tracking and I can actually do application match an implicit applications like there's a printable sticker that slaps on the side of an Amazon box and that sticker goes when that box is set on your front porch because it since that your cable modem was the closest and it seized your neighbors around as well right and now that's what that's a game-changer right now we have something that is is a completely different model in terms of you know enabler in terms of what's what's possible and that's just one example of the sensors not about the wireless and the technology to get it there and I can you know spend six more hours on that because that's part of you know passion and that's what we we do but that's not that's only part of the story right what we're really doing then when we now we could do these sensors everywhere and now we can gather all sorts of information from the world around us what really matters is so what we can do with that you know what sense are we gonna make them what what are the applications we can do on top of that and how many of you are just a little bit creeped out sometimes by the ABS you see right yeah at dinner last night my wife was telling me that she saw an ad for something that I had been searching for and told her about but I hadn't even come home yet so normally when this happens I assume it's because we're both using the same internet connection out of our house and caching the servers that Comcast are basically joining our searches together and this is fairly common but last night she was yesterday afternoon she started seeing ads for things that and it was a noticeable change because I was doing some kitchen renovation that this was something I was searching for she saw an ad for it something I had searched for an hour before not anywhere to your home the only communication we had was was on the phone and now we're really creeped out so yes this is this is it's amazing what's possible and kind of scary too all right so when we talk about your big data all the time was they did you know so I wanted to take a moment to define the term you know we have a whole research that we have research centers at Georgia Tech on big data we have master's degrees and analytics and all these sorts of things happening now but let's set the stage and make sure we all agree on what's Big Data and what are the challenges and IOT is contributing massively to this right so we can't talk about IOT about talking about big data so the first week we describe big data in five feeds right the first one is volume lots of data right that's the easy one it's coming at us and seeing a mass right not necessarily big chunks lots of little chunks sometimes but there's lots of it it's data coming from a variety of sources all kinds of sensors but not just all those sensors what else apparently my web searches for kitchen appliances apparently my car service conversations about kitchens appliances right there's all bunch of unstructured data from different sources that's all coming together and and part of this big data story when we talk about the analytics on big data one of the big challenges is this variety and the fact that so much of it is unstructured it's not just a lot of it but it's coming fast so with high velocity and then those are the three B's that I put my group in the how the data is coming at us the other two piece are more on how we're dealing with that data and how we're processing that data first is the goal of big data analytics is to bring us to learn something from them to bring us valuable insight right to make the decision that someone else at Russell's house must be interested in kitchen appliances I'm gonna be hung up on that story but more more importantly to gather the inside from all of this information about weather and traffic and sports activity schedules and whatever else might affect hey we should adjust our maybe we should adjust our urban delivery you know our trek rolls so that they avoid I don't know national championship football games happening the same day they think it's going to snow and you know whatever else is going on in it later at the same time so when we talk about you know optimizing you know shipping or delivery or even you know how how we package things right now we have this whole wealth of possible sources of information that we're going to want to bring to the table but it's something useful you know something meaningful we've wasted our money on all of this unless we can actually come to some conclusion that improves we're doing so it improves our delivery schedules or whatever and then with this is the one I'm you know they stretch to get another beat right they say God what's a fee word that tells us it's a voracity the point is this one is actually one of the places that I think the research and the sensors in particular this is actually the fun stuff where it gets interesting my best student projects that you know that I have students working on are not the ones with the cans clean sanitized data set that they use in every database class every semester that just you know a big chunk of did no no the good projects are the ones where it's real sensor data and two-thirds of it as noise but I don't know which is the noise of which is the good stuff and we got to figure out which of those values we can trust and which ones we gotta throw and that is really where the you know that's when to me some of the most interesting work is some where some of the most interesting applications of AI and machine learning and all these interesting sort of analytics techniques that we work on we're trying to figure out you know which of this data is good at which of it is not before we can go start gathering you know making making decisions about all right so some are 85 bees have been data now you know there's your you know we can you can win an argument cocktail party now over what's the definition of the big day alright so the other thing that when we talk about you know why today why do we think this is possible the other thing we have to leverage and the other thing we're doing that makes IMT applications in the real world feasible today is the clap look by the cloud I mean the fact that we don't have to open all the physical compute and storage resources ourselves we get to leverage those from third-party providers we get to allocate them quickly and then deallocate them when we're done so that we don't have to pay for resources we're not using and without that sort of capability you know dealing with this huge volume of data would be much more expensive so this is a big part of this as well as we moved in to more applications of biology and this is really in the really comes to play with smart cities work where we're talking about deploying these sensors and technologies and the urban environment where we're going to improve traffic over to improve the way the city functions more and more what we're finding in this space is the importance of data sharing agreements the fact that no single place you know we want to do this big data analytics we want to bring in all this information in order to make better decisions we can't do that by declaring okay one person is going to own all the data one entity is going to learn all the data and that's how we're going to get it done there will always be some other piece of data we need that we want to add to it and the there will always be some need for an agreement to figure out how to get to that right and so these types of the Greek what I mean by agreement well we have to agree on lots of things with their data formats are sort of the you know the easy one you know how we're going to connect and move the data but the actual harder ones are what about data use policy what am I allowed to use the data for liability has turned out to duck to be a road block for some of this data sharing whether we're talking about cities sharing data with state governments or between cities or in the supply chain space and we're talking about in curbing logistics for instance with urban delivery guess what we've got a whole bunch of competitors all trying to make deliveries in the same space at the same time and in order for that optimization right to really work we actually have to we can't just you know focus on the UPS trucks that might wvs but it doesn't help them you know to some degree but it doesn't help the city as a whole or the solves a problem is we're going to work with all of them and it's not just a big shippers thanks but all the smaller ones as well right and so the the space is a challenge from the point of view of you know people protecting their own interests and their own you know proprietary information but then also the liability piece the easiest examples a liability one is you know yes there's lots of interesting things we can do about optimizing response time for ambulances right but what happens when the data we relied on to get the ambulance of the right place at the right time actually was some of that 30% of data or 60% of data I should have thrown out because it was wrong and we actually since the in goes to the wrong place or they were delayed in it for 20 minutes or whatever it is right when someone dies because my really smart algorithm actually made things worse now all right and so even though when you start the conversation of data sharing everybody says oh yes we should all share data when you actually get to the legal agreements that we have to sign somewhere there's going to be an indemnity clause right that's the language we use that says you know I'm not responsible for anything that's in this data and whatever you do with it is not my problem well those are actually really hard things to get you know especially a guy that works for a state institution some of you have dealt with contracting at Georgia Tech and in many classes are kind of those things that are really hard to get good sign though I have a whole bunch of data sharing agreements that are stuck on this this issue right for instance so anyway for good buddy another talk nobody said hey what about walk to you yeah I know it blockchain solves everything right we'll talk about that another time but this data sharing is in fact the potential applications there's some interesting work in the use of blockchain for data sure all right so to summarize this sort of overview of technology and big data and where we can apply them what obviously we you know I'm here because we think there and we have seen lots of opportunities and started exploring lots of opportunities in applying IOT and big data to to supply chain logistics asset tracking all sorts of infrastructure monitoring and maintenance you know predictive maintenance of you know I want sensors on every piece of equipment right you know vibration sensors on motors and pumps and the gears houses whatever so that I know I can so that I can prepare to do maintenance on it instead of waiting for the thing to fail tragically ended up down for two weeks right those are the kinds of real applications that that we're seeing storage sensing obviously well the whole cold chain right space anytime we have to worry about climate control in to in throughout the process we're seeing lots of application of an interest in IOT for that you know alert the driver of that though how many of you been driving down the road the refrigerator and behind the refrigerator trailer blowing water heaven because clearly that cooler broke and this thing is melting right someone should stop the driver and say hey you know hello cream fix the refrigerator and so anyway security monitor your customs processing multi-modal rounding multiparty data sharing as I talked about and then what's near and dear to our hearts from the point of view of IOT and supply chain this is the education apartment all right so just some a few slides with some stats on you know it is it's Russ making this offer is this business real and the reality is that we are seeing a very real spending on on connected logistics solutions globally and it's and it's ramping up just one example of fleet management right so commercial fleet commercial trucks right think delivery vans everything from the the or you know installers and maintenance crews as well but everything from the small FedEx Dan all the way up to the over the road the interstate truck long-haul truck in the last you know it was for several years they were there was pretty good investment in pretty good adoption especially from the big carriers right you know the the freight waves are the Tigers or the Jamie humps or something like that but now we sort of that adoption threshold to where basically you can't be in the trucking business and not do this you can't afford to be able to avoid doing this and we're seeing it you know all the way out to the sound to the even smaller operators alright so with the time that's left what I want to do is talk about some projects that you know I mentioned this this IOT thing it's just the two days for him you know next year I'll come back and I'll call something else meaning to your time but the point is you know this is the kind of stuff that the Georgia Tech has been involved in in leading for many years and these are some of the projects that we've been doing so one of the focuses we've had one of the for many years is this issue of tracking without GPS so the location determination without the GPS why wealth because in this room the GPS receiver doesn't work in general indoors GPS doesn't work work but it's not just indoors you know what that urban Canyon is we heard the term Urban Katee what's an urban can if there were a cut like here you know in the Grand Canyon when you're driving down the street in the big city because you can't see the Sun because it's blocked up from all the buildings right basically if you can't see the Sun guess what you can't see the GPS satellites either so you know driving down a busy street than city you know in a congested city typically where you know waves of trying to rewrap you because you're stuck in traffic well guess what now waste thinks you're three blocks away because it doesn't your GPS isn't working either so lots of ways that we can address this issue and have looked at it over the years using other location infrastructure I mentioned this use case for Laura dense deployment and being able to use the lower base stations to detect what is the location of the sensor again without GPS in that case it's not so much and so there's really two advantages one is you know maybe I can't take the GPS but the other advantage is GPS there's actually still a really expensive chip to operate even though it's just a receiver on a transmitter it's actually consumes a lot of power to power that GPS and calculate your location so the real advantage for this being able to do this without GPB ship is remember that printable sticker oh and slap on the side I don't want it to cost have the costume associated with the GPS receiver but I also don't want to have the battery consumption associated with GPS receivers so all sorts of ways to think about this Oh location versus proximity a lot of times people say I need to determine the exact location of this thing because I want to sell how close it is to this other thing well actually then if your application is how close am I to this other thing there's a lot of cheap there's lot cheaper ways to figure that out then if then determining your precise location so proximity how close am I to that door how close the money to the destination that I'm trying to reach that that shelf that I'm supposed to be sitting on whatever it is and so we've been working in the space with test beds across campus for many years and different versions of it and including the the campus Wi-Fi infrastructure and we have api's for determining your location the location of a Wi-Fi device based on where it is relative to the infrastructure but what if the question is not where a mom but how many people are around me or how many packages are around me how many devices or how many or what if it's not around me what if I'm managing the campus facilities and I want to know the utilization of different classrooms on campus right or I want to know how many people are using this how many people are waiting in line at Starbucks is actually one popular question and so it turns out that the same sort of infrastructure that we've developed for determining location if you turn the question around and say ah I just want to know how many people are in this space that same technology and many of those things the same tricks if you will can are very good at doing that and so what we've done basically is develop a set of api's for the again leveraging the campus wireless infrastructure but basically using you know this device that we all carry around as a proxy for a person right how many we're gonna we're going to basically use the number of Wi-Fi connected devices in this space as an estimator for how many people are in this room and so we've actually done studies and what's interesting is what do you think the number is what's the ratio weight devices per person at Georgia Tech the device is connected to the in the radio 2.2 yeah so when we started doing this work I don't know 18 years ago it was pointing right and now it's more this doing more than two right and what's crazy is you know every couple of years we have to go back and validate try to be able to send people out with clipboards to actually write down have them count the people in his place and then go compare it to our API because it you know it keeps changing in you know I think I think I've got at least three all right now all right and so it's actually a very interesting study in itself as to what's this ratio but the point is this estimator turns out to be really good and it allows us to do things like so this is in cloth counting the number of people in the space around Starbucks so we can't tell if you're in line at Starbucks or not but we can tell if you're in fact you know within say 30 feet of the line of Starbucks right and the first thing you see is this so the yellow line is Monday Wednesday Friday and the red line is weekends and then the first thing you see is this pattern where it's obviously class changed could people close their laptops they get up and they move around and then they turn it back on right and what's anything a place like Georgia Tech of course is that picture looks different Monday Wednesday Friday versus Tuesday Thursday because we have a different class schedule right but we actually use this in the early days after club open right to help motivate Starbucks to stay open past midnight because originally they were closing in like 11 and favorite we said you know look there's people out the building until 2:00 a.m. there's you know there's a big group of people in the building till 2:00 a.m. and then there's a pretty good-sized group after them all right and so Starla you know open later and and what else can we do it's not just happening people in his base but what what are the mobility matters what else can we do with this type of infrastructure is look at where things move around campus right where are these radios move I don't have to know who the people I just see these radios moving around and I can make a pretty good guess that both of those radios are associated with a person and that we can we can actually look at movement around campus over time all right so this we work with facilities to help young study things like we just did a major change of the way skipped class scheduling does works on campus right we finally have a big enough gap between classes in between classes this semester all of that is is you know those kinds of studies and those kinds of changes are Kimmy can be document it can be motivated by hey we've got real data now to show people are actually trying to get from Iowa to college of management between classes right this isn't what nominally lots of people actually trying to do it i I have a class schedule where I had to do it teaching it wasn't any fun that I was late every time right and it's much easier now so ah transportation needs the the Tech Express show right we have data up showing exactly that yes there's a need for this lots of people moving between club since x-squared directly right and we can look at how turning on you know how deploying that shuttle actually changed people's patterns people actually would hit there and get on the bus rather than walking the other way parking needs new walkways all sorts of interesting applications of that alright so back to the lorwyn discussion one of the things wrecks and tying it to campus what we're actually doing is this pointing the space station technology on campus to use so we have some devices that are similar to this I guess one of them you can see mounted to the outside of the aware home how many of you heard of the aware home son yes it's a house the church their homes purpose-built for doing in home studies technology application in the home for instance and so it also turns out it's a nice sort of on the edge of campus location for us to put one of these and then test okay how you know how far at the home mark how much of campus can we can we reach from there and the idea is you know to be looking at the you know what kind of density do we need and what kind of applications can we do at different densities but basically we're deploying this on campus and it's open for the rest of you to use right anybody that wants to play with do some sensor applications some studies on campus we're setting ourselves up to support that and that's what we're trying to that's what we want you to be able to take advantage of so far we've done some some capacity testing both mobile and stationary and what we're expecting it is with with two or three of these base stations we can cover all of campus including inside those right that's part of all be out mm all right because this is really low power level data ready right and it doesn't actually have to you know when I'm trying to watch YouTube on this alright and so I don't know maybe it's six maybe it's 10 maybe we got double the budget and deployed 20 but for Wi-Fi we've got over 6000 access so we've got one two three probably four in this room the cloth classroom is the big ones have like 18 access points in it so you compare the infrastructure costs they would start the deployment to a technology like Wi-Fi and it doesn't even come close all right similar my cellular right even at 20 even at 10 nodes we will have fewer Laura base stations gateways on campus then we have cellular sites deployed by AT&T and Verizon right on the top of several buildings and inside several buildings we have cell sites on this campus because we gotta get that that infrastructure in a bandwidth guy all the way into this kind of space right a very different scale of deployment again why am I talking about this as being viable for supply chain today this is why right the infrastructure is so much cheaper and is so much more feasible to deploy at a scale that yeah for a few thousand dollars I can cover the whole shipping armor right I can cover I can put one of these base stations on every ship right and half of you know one cellular or satellite uplink but the tracking of every package inside all right and that's the model that we use for thinking about this so some projects that are going on our campus and I'll wrap up here with just a few more slides working with Matt Swartz at Center for GIS on they have this long-running project of going there with all these sensors and what they've been typically using Wi-Fi what we're doing at this point is is planning to augment those sensors with Laura so that we have what it's going to give us is a really nice reference point right so we're gonna have this device we can get to and configure advantage because it already has a Wi-Fi but we can we can add the Laura capability to it and again as part of this this proof of concept and testing to figure out what what Laura is good for and what it takes to really maintain urban beehive have you seen the you know has bees right yeah very cool work one of the interesting applications here is how do we know how many bees are in the in the hive or how much honey they've produced well it turns out that's actually pretty easy to do just by knowing the weight of the heart and so all we really need is once a day a few times a day get a data point from that thing wherever it is you don't need any wired infrastructure all you need is a sensor on there that will tell you a few times a day how heavy is this thing right and so perfect application while you're at it Charlotte imagery the vibration or whatever else and we'll see what we can come up with right and so a neat project that that again sort of proving out what's possible other sort of things think of any place that you might want to know how this thing works or this data better the health of it but to date it's been impractical to run to get real-time data because it was too expensive to run wires in power you know network in power too so instead you know there's a sunscreen dispenser in a park all we need to know is is it still working and how much product is left in it a couple of couple of samples a day and for you know just a few dollars added to the cost of this thing we can we can get that I I would be remiss in talking about these things without sort of talking about how I got here and so what connected me and my work to the space and the supply chain group and that's this long-running project we we've had with Professor Bartoli nicely monitoring or augmenting not just monitoring but to improve the transportation system for bus systems and in particular focusing on the bus budging problem so nothing worse than walked watching two trolleys follow each other right around the campus and then the worst part of that is you know 20 minutes before you see the next one and so trying to improve on that and that's kind of the at how I as the IT guy and the networking guy sort of started thinking about these these supply chain related problems and logistics related problems over the years I'm going to leave you with this and say there's lots of opportunity here and I'm excited about it and I would be happy to talk to you about projects we can work on and things you might think might be interesting to try it out on campus and you know if you have an idea for actually you know playing with something to do a trial on campus for a class project or a research project or just some crazy thing you have for a startup I mean right that's why bill and I are here but there are some things we have to worry though and there's some things that we obviously have to pay attention to as we bring the entire everything online privacy being a big one both personal privacy right about us and tracking location and knowing too much about us but also this this data sharing issue among competitors there's a lot of work to be done there and sort of navigating the space of yes how do we actually do the you know rising tide floats all boobs but in order to get the rising tide we actually have to all invest in something you'd be willing to give something out well actually go possibly the we think is proprietary and our competitive information and there are some challenges to that security I think the core one of the core issues with security and IOT right now is the race to the bottom on priced right and so you can get for instance a a really cheap webcam now right for ten bucks to put it on the Internet well guess what why did it only cost ten bucks because they're using a free software stack does no one have done security updates in 20 years and guess what as soon as you put that thing on the Internet it's going to be owned and it's going to be part of the box and it's not just going to be your your best hope is that they're not transmitting pictures of your house and sharing them with the world there's actually a website you can go to and see all the open webcams and I don't post it so you'll have to Google but then you know that's probably not what they've stolen your phone for what if they stolen you're coming your your webcam board or your smoke detector or your whatever they've taken it over so they can run the Bitcoin you know they're there they're looking for processes either in the world but they can harvest Bitcoin right and so anyway so the point is that the challenge is that you know hopefully we decided we found the bottom we're going to bounce and we're going to start taking the investment in the security of the software both hardware and software security of these devices more seriously and say look everything that comes into my shop you know every devices connect to the Internet in my warehouse there's a potential liability and we need to make sure we invest appropriately to secure it and manage that liability similarly that leads me to the maintainability aspect I have a whole bunch of commodity IOT home automation devices in my house that have this you know decent out what we call the out of the box experience where you you open it up and you're all excited and you follow the in set of instructions and you get it working and you get the app and you can make the light turn on and off from across the room with your app and you go wow that's cool and then it doesn't work one day you have no idea how to troubleshoot you have no idea how to start even fixing and so typically what I do is I go back to the start of instructions and I push the reset button I start to the you know try again and if that doesn't work I throw it away because there is no help for how you debugging right there is no no solution to fix it and then you know one of the things we talk about that's a big issue in this space is the what I call the lifecycle mismatch right how many of you have seen apps or purchased your very own LED light bulbs right replacement LED bulbs right how long do they how long are they supposed to last years right I I purchased them that same 20 year life expectancy right and when we talk about especially decima in infrastructure right street lights you know realigning my entire warehouse right the investment the expectation of the people building something like that is this should last 20 years right I I can't I can't afford to do this every year all right one of the big problems with that yes can you name a radio technology that has lasted 20 years how how long has this device using Bill Self little smart all right I said the smartphone idea is 10 years old right but guess what the radios on here are about six years the radio that was in I have my original iPhone because it's cool right it's a museum it goes in my museum and my kids go oh my gosh it was so thick heaven let's introduce them and it still powers on and it still works but I can't make calls with why because AT&T turned off that was a 2g radio AT&T turned that off four years ago doesn't work anymore they turned it off because they needed the spectrum to support 3G and 4G and all these new things were doing it makes sense more efficient use but I can't deploy the streetlight with a 20-year led lifespan clean and even pretend that that radio is guess what the Wi-Fi radios if you have an 802 11 B Hoadley ratio in your device guess what we don't support it on campus anymore we turned it off several years ago from the same reason that AT&T turned off 2g by supporting your 10 year old radio it slows down the rest of us all right doing with you know everybody else so we turn it off and most people that operate Wi-Fi networks do the same thing so yeah all right now where we see this issue again where you know life lifecycle on from this this radio infrastructure it's still much shorter and the refresh rate is much shorter than the these other technologies we're now attaching them to write my my Wi-Fi thermostat is another good example all right when I put it on the wall I replaced the perfectly good analog thermostat that was still working and was thirty years old how long do you think by Wi-Fi thermostat war his work at best in ten years it will still allow me to walk up to interchange the temperature I'm sure the Apple working honey what we'll have decided they don't support it but I'm sure I mean that's my best cases I can walk it that it will work as good as analog one mentor gospel right so I didn't throw anything along when it's still in the box in the closet so that when I come in one night and the internet is down and I can't turn on the heat I'm gonna rip that sucker off the hook with the elevator right so and then finally and then you know that leads to this resilience issue right the whole point of this is to improve resilience I think that's probably our our 2018 buzzword applied to the IOT spaces resilience you know being able to to respond to disaster and deal with the unexpected and our mantra should be as we deploy these things we're going to increase resilience not decrease it right and to many of the technologies we deploy become a new single point of failure and actually reduce our resilience all right thank you for your time and the gun thing a few questions are known as toggle house okay thoughts question is discussion is well how can you use this up in your life what do you love what do you what are you excited about and what are you afraid of [Music] yes great presentation two questions you delve into what do you think Georgia Tech could distinguish ourselves so so he he led the question with was how do I think Georgia Tech can distinguish themselves in the space relative to the MIT s and the Stanford's and obviously I'm biased and I think we've already done that but the XO I think and you know honestly this is why I'm excited about this work in the space is because we we have a well-established reputation obviously top program in supply chain around the world everywhere I go you know people know that and refer to that and but we also have these the the ability to do this kind of deployment in real test but in an urban environment right there urban logistics is one of the sort of key channel the next generation of you know moving every more more people into the bigger and bigger cities and guess what we happen to have a campus that sits in an urban environment the campus has a really nice testbed but then we don't have to go very far yes you know this this urban Canyon problem isn't just a you know a lab exercise for us I have it every time the trolley pulls in front of the MARTA station it's blocked by three new buildings that it can't see right and so congestion dealing with you know these day-to-day realities of life and in the city Georgia Tech is is a perfect place to do that kind of work and excited about some of the stuff we can do here good question yes Russ and you got great presentation are you doing anything with the north yeah we not day-to-day directly absolutely directly involved in and one of the application one of the aspects of that is this sensor package there that's being deployed prototyped and deployed on the light poles across you know along the North African corridor and we definitely work with that team and you know very interested in you know it brings up a lot of these issues the challenge of what should be in that package and how you maintain it and I deploy it and maintain it and support it but also the data sharing agreements right we're we're already into the oh we're going to put cameras in there who can see that David right what is that right and who is and how do we make that data how do we make data available from these things and and deal with the privacy and security but also the scale bill right one of the big impediments we run into with data sharing is oh I would love to share I have all of this real-time data from our transit system and we use it every day this is available and it's very reliable but I can't let you have access to it because you'll crash my server alright and so working with them to scale that right delivery of that data is a big part of what we we have worked on in the city space yeah good question other thoughts questions yes projects so we have not gotten our hands on any of them yet the front flip from the Laura and the Laura space so we've done we've done RFID stuffing like that in the past but these new printable trackers are just announced like three months ago or something I think yeah I mean the the plan has been there for a couple of years we finally have seen that someone's actually starting to manufacture these so we're gonna try to get our hands on something as soon as possible and start doing some of that testing and what I mean ultimately there's there's a bunch of work to be done and you know now you know we we understand the theory how will it work in practice especially in the urban environment where guess what this you know spectrum is limited and these aren't the only radios we're going to see right we're gonna see there's gonna be a lot of noise and so one of the the research areas is really in you know how do we optimize for battery life and you know sort of minimizing cost but at the same time deal with the noise we're going to see in this radio space and and what is that what are the requirements in terms of the actual density of those base station for those gateways right we we know we can do six or eight or ten and get coverage across so that I can get a message three times a day that says the temperature of the package right but if I actually want to tell you the packet just got set on my front porch that's gonna require a lot more density we don't know what that density is yet that's so that's kind of I think though so that's one of the sort of research questions if you will is is figuring out for those different applications we might be interested in asset tracking what are the trade-offs going to be power cost and density of the base station but love to talk more about it and look for again I think figuring out in you know supply chain is a big space where housing is an X base you know let's just six is a big space there but there are specific applications where we can say ah let's focus on this particular question of tracking within a certain type of warehouse or a certain type of manufacturing process or whatever it is yes you might mention you but how much is the cost of those they just put that piece of hardware yeah okay so this is an interior gateway cost is around $500 101 something that you're going to mount that Wars like we have a deal where mountain is more than a $1,200 range and they're there they're coming more for supplies in there was the price a little bit already so these different in that we have in a prototyping the goal for the printables that one sticker version is well under $1 probably years out from that functionally there's really no difference between those sticker based ones and one of these except the batteries are much smaller and you're really only looking at that battery to last for whatever the duration of the shipment is so we're talking about over a week or two and in that case you can do a 3d printed yeah and these you know these experimental kids have USB port so in your program obviously they're talking the USB port on the sticker so I mean think about the scale all right the biggest challenge is assuring that I sharing platform so what you need two things that are essential for it to be effective could be brightest in one hand and on the other one would be transparency because you don't want to share some information that our added value for in business and on the other side you want to make sure that everyone else is getting the right information from that edge as well so you need my kind of an administrator at some point could be a government entity and they need to have transparency otherwise but why would they do that when they don't want to use this particular added value that the ABA's information and put it there so I don't know that can happen and yesterday where no I think it wasn't man that while he was talking about smart city and to other smart city you need to have all the information not like three entities but all of them and it seemed so complicated to me especially in a capitalistic system that yeah I completely agree I don't think I have an easy answer for you except you know part of the part of my goal with with this when I do these presentations is absolutely to raise it and say you know take to your points that we have to address this and and the only thing way I know you know we talk about you know you can't blow the ocean I'm not gonna solve them all at once we're gonna we're gonna have to find some specific examples where you know people will agree to share and then see value out of it and remarkably actually in some of the related but you know I should say unrelated work but with some of the same issues we've had on with one of our other hats on is in healthcare data alright so similar they're similar very serious challenges around privacy and and competitive environment but there is a greater good of improving the way healthcare operates right and so we are starting to see and you know we have examples of successes and moving that beast forward by the Machine forward and to doing some better data sharing I I think that you know and but by no means are all the issues solved and we have to keep fighting it you know sort of one new project at a time and one new application at a time and you're exactly right the as we as we've done this around the you know work with projects around the world you know every culture every business environment has a different set of constraint we have to you know do something to start some of this over I know some of the things that are possible to do here I can't even begin to propose in Panama on the other hand there are some things we can do there because it is a smaller more well-defined space with you know you can go down the street and literally bump into government officials right of the central government and that makes things easier to solve than trying to solve it at a scale of up here so anything so you're you're dead on you're absolutely right like for instance in friends with a very different yeah so it could work right there with you can share information everyone benefits right into bands by can I ask you you know the impact of all those cunning cheap sensors that we we are always connected were receding waves and long-term what's the airbag Oh oh heavens that's a totally different conversation no yeah Wow we could have a totally different well I'm not no expert in that space at all other songs good questions yes we're having the trash we want to put sensors everywhere right well big of a challenge is or security concerns since I'm tracking everything that means our data and security I can't even begin to estimate let's just say it's real in this big right I mean the you're asking you know what's the cost of security no I mean I think it you know it's in my pitfalls there's two of the three things I think it might it's all solicitor things you mentioned there right I think so there's absolutely real issues with securing the devices and you know being able to maintain the infrastructure keep it protected but then what about the data and what it reveals and especially all the data complying what it reveals about us personally and I think we're going to have you know we have a long way to go in terms of we're going to see you know regulatory ask you know aspects of this we're going to see liability right you know I think you know the lawsuits are just beginning in terms of what people being exposed what's but perceived liability proceeding risk is in terms you know we don't even know you know so now that this big conversation tuk-tuk things to work on but hey that's why we're here to solve these problems all right thank you all for your time man

Channel: GTSCL

Views: 14,053

Rating: 4.8775511 out of 5

Keywords: logistics, supply chain, IOT, internet of things, georgia tech

Id: hx1ap5yPMos

Length: 78min 49sec (4729 seconds)

Published: Mon Feb 12 2018