 Just introduce Emily King. Emily is the Program Manager for Research and Extension in AWI. Emily hails from the Central New South Wales and has always worked in agriculture, beginning her career with Australian Livestock and Property Agents Association before moving to AWI in 2012. Emily has worked in a number of teams at AWI, but currently fills the role of Program Manager, Research and Extension, and is responsible for the reproduction and nutrition research portfolio and grower extension development. Today, Emily will speak on smart tags, what are they and how do they work. She'll then give an update on the latest research and how producers could use these tags to monitor animals and improve management. So, thanks very much, Emily. Thank you, Emily, very much for your kind introduction. There's that ram from before that everyone liked so much, just for you to feel. You'll notice probably that he's a bit messier on the nose in the video. He was in the still picture earlier. So, thank you very much for having me here today. I will be having a bit of chat about smart tags and some work that AWI is doing in development there. I'll also have a chat to you about some of the work Henry's already covered, shearing shed design and what we've been doing there, but we're also doing some work on technology, wearable technology for shearers so that we can try and reduce the fatigue issues and stress on the muscles and that sort of thing for shearers so that we can hopefully keep more good shearers in the industry for longer and improve their work life and also reduce the risk of injury and also a little bit on wool cue as well today. So, Henry, obviously you've already heard from and just down the back of the room we've got AWI's chief operating officer, John Roberts, who's actually SA base. So, he'll be here today if you wanna have a chat with him and Andrew Dennis as well, who is the SA wool cue rep. So, I will crack into this now and looking at AWI smart tags to begin with. So, you can see there in the ear of that you there, there's a tag, a white tag there. So, that is this tag here. So, these are the tags that we've been working on. So, that in the front of that, that's a solar panel. So, we've deliberately chosen not to go with GPS in terms of location. For these I'm using triangulation and using low range RFID for this. So, there's a couple of reasons we decided not to go with GPS. One of them is cost that that would put into it and then the weight of the component tree as well and also the power demand that they have. So, with this solar panel, we can keep that running with no battery, no nothing in there. We can keep that running on the sheet to provide the power to run the tag that we need. Then, you can see on that lamp there that there is a collar on the lamp. And so, these are the collars that we've got. So, this is stretchy collar. So, you can use that for things like mothering up. So, that you would wear a permanent tag and you might just collar the lamp for a short term period to do the mothering up by proximity, that sort of thing. And then, I don't have one here today, but in the bottom right there, you can see a harness on a ramp. So, we're looking at some reproduction work at the moment that I'll explain a bit more about. So, as I said, the AWI Smart Tags that we're looking at are solar-powered ear tag and there's also battery collar and the RAM harness. And we're trying to make these so that they're a combination of different technologies so that they're not just, I guess, a one-trick pony for one of them better phrase. So, we're looking in there at having an accelerometer. And so, the accelerometer is able to tell us things about whether the animal is walking or grazing or ruminating, all those types of things. Location, so location of where they are in the paddock and proximity, so proximity of one sheet to another via the information that's fed back by the tag. So, that becomes really important if you're trying to have a look at mounting events for joining or whether the animal is grouped up in a really tight, whether they've been pushed up in the corner, for example, if there's a predator in the paddock and there's some abnormal behavior happening in the paddock. So, to get that data, we use these receivers. So, these sit out in the paddock. This one's a bit funky and yellow looking because it's been out in the paddock. So, these we use, we just, at the moment, just attach them to a star post in the paddock. And then the range of each of these boxes is about one kilometer at this stage. We are trying to improve that and so that's over fairly flat sort of ground at the moment if you had, if you've got more hilly terrain, it won't cover as far, but we're trying to improve that all round. And as because all of the location and stuff works on triangulation, so you need signals to bounce off a minimum of three of these boxes out in the paddock to understand how far they are from each of the boxes to give you the location of that animal out in the paddock. So, every day there's more than 45,000 data points collected per animal that's wearing a tag. And so, obviously we don't expect you to actually sit there and find useful, meaningful information in 45,000 data points per sheep per day. So, what we're also doing is writing algorithms. So, I guess, computer, getting a computer to understand what all of those bits of information mean so that we can just spit you back. A quick message or an email to just say, oh, think there might be a problem. None of your sheep have gone to water today, potentially. Things like that. So, a few of those things that we are looking at, we're looking at welfare alerts. So, for example, can we understand by the movement that the sheep is making that they have a high worm burden without you having seen it yet? Is there a potential to pick up when sheep are lousy? Is there a special way that the sheep moves that we will be able to see, for example, can we pick up that they're always rubbing up fences or trees? Grazing and feeding intake, having a look at net feed efficiency and how sheep are utilizing the paddock and the regions of the paddock in which they're grazing. Predation alerts, so, you know, could be anything but likely a dog or a fox, I guess. And reproductive management as well. So, this property here that you can see in the picture, that's a property at Buck Holden in central Queensland. It's about 28,000 hectares. So, we're looking at it as one of the trial sites so that we can get some extensive pastoral zone testing happening. And so, the work there is being done by AWI and Central Queensland University, having a look at whether we can, in fact, pick up predation and the disease for the worms as well. Hasn't been particularly wormy out in central Queensland in the last little while. Hasn't been a great deal of rain. Hasn't been a lot to pick up there, which is great on one hand, but if you're trying to pick it up, it's not real fresh hot. There have been some challenges there in terms of robustness and retention for the tags. So, we're working on that at the moment, and then we'll be putting those tags back out, but this is why we're doing these trials so that we can pick those issues up before we go to market with that. Grazing bites is in collaboration with Murdoch University, which is at Perth. And you can see there that this sheep is rigged up with a poo collection bag there. And so, what we're doing is collecting all the fecal matter that comes out of that sheep because we've got tags in and so we can understand what the sheep is taking in versus what is coming out the back end, or isn't coming out the back end in some cases. And we're collecting this data across a range of grazing types. So, mixed sheep cropping and also out to the pasture zone and the high rainfall zone so we can understand across different types. We're hoping to accurately predict feed on offer by understanding the grazing behaviour of the sheep, how quickly they're walking and moving through the paddock, all those types of things. At the moment, the baseline early stage data is telling us that we can't accurately predict at the moment down to one day, but we can get it pretty accurate out when we take in three days of data. And we think that at the moment, we might only need to tag 40% of the mob to get accurate readings for the whole mob. The reproductive projects are being run in collaboration with the University of Sydney and so we want to understand how we can better predict and understand reproductive behaviour. So we're looking to understand, you know, detection of estrus, having a look at mounting, you know, for example, we might be able to tell very quickly if a ram has only done a quarter of the work that the other rams in the paddock have done and whether his libido is a bit low or something like that. And also lambing events. So, I mean, you know, some applications for this potentially, if we can easily detect estrus, we might be able to have, you know, a hormone-free AI protocol, for example. So this is a bit of a video and you'll see these spikes in the graph here are the mounting events. And so you'll see, so this ram that's about to run along the back of the shed there is tagged and so you'll see what happens here when he runs through. You thought you were going to see some ram fall on today, didn't you? No, no, no. So we have here, so we can... This trial was, as you saw, on a research site and it came back, it was quite good. We were able to quite well pick up the mounting events and all that sort of thing. When we put it out into its first paddock trial, the results weren't so clear. So sometimes there'd be other rams hanging around, so the proximity was being affected, all that sort of thing, and other youths hanging around. So what we have done with that. And so, yeah, so we've got that trial running at the moment. We did, because we wanted to make sure all of those animals came in on time, we did cedar those sheep to bring all of their estrus all synced up. And so then, so that, you know, basically when the researchers were out visually checking mating times, so that we could match it up with the data. So that trial is underway now and all of those youths should have been joined by now. So one example of something that actually happened with a flock that we had the smart tags on, the manager of that place rang up and said one of the youths that had a tag in has died. And he thought that might have been wild dogs that had got that U. So we went back and had a look. And so this was in the early days when we were still collecting data and we didn't actually have the algorithms in place to understand what the data was saying in real time. And so we had a look here. And so that line of orange bars across the top there is when she was grazing. And this is her just standing around and the bottom is walking. So you can see she's basically not walked at all. So this time period from here is May 10th through till the 20th of May. So in 10 days, she's barely walked at all. She's done a hell of a lot of standing around and a little bit of grazing just in her immediate area. And so we actually went and had a look at that U. And she had twins on board. So we concluded that actually she probably more likely had died of pregnancy toxemia than had died of a dog attack. And so this predation that you can see here is post-death predation. And you can see as well that there was a long period where she didn't move a great deal and then there was a little bit of head movement there and we think that was bird predation on the carcass that has moved that ear tag around a bit. Also this graph here is showing you the closest tags to that U. So what she were in proximity with her at any given time on any day and for how long each they spent. And so here you can see that probably what we'd consider to be a fairly normal amount of contact with the other sheep. And then on the day she died, that spiked back up again. So it looks as though a lot of the other sheep came in, had a look, checked out what was going on and then they're basically after they've realized she's dead have completely left her alone and haven't gone back to the era. So we think that there's also some interesting things that this type of data can tell us about behavior and about flock behavior and how animals move in family packs and that sort of thing as well that can help us better understand how they operate. Shearing is one of the most strenuous jobs out there and is one of the ones that WorkCover has the most injuries associated with. And so what we want to understand is could we put some sort of sensor onto a shearers body to understand when they're fatiguing and see if we could maybe give them an alert or something to say, you know, you're getting into a period of high risk of injury, you need to stop or you need to stretch or you need to whatever case may be. And so in this work, when we're trying to understand shearers injuries and how these are happening, this graph basically looks at the muscles in the lower back and how well they stretch and then recover. So the higher the dots on here, the better the recovery. So you can see in the first run, things are going quite well. I mean, there's a fair bit down in this bottom section, but the shearers are recovering quite well. And as we move through the runs of the day over out to run four, he's basically got nothing up in this recovery. So all he's doing is just continuously straining his muscle. So what happens is when those muscles stretch, they need to retract back in the lower back and it just wasn't happening. They were just staying stretched. And so then when, for example, a lot of injuries happen during the catch and drag and usually due to a quick sudden movement. So for example, if a shearer goes in and goes to tip a sheep over and then the sheep sort of ducks my mind, they quickly reach out to grab it back, then they can put extra pressure on their back and that can move. And their muscles aren't taught enough and snapped back into place and resilient enough to hold all of that back together and things can slip out where they shouldn't be. So what we wanted to do is look at, is there a way that we can have an alert or try and mitigate that before that injury of risk occurs? So this is a video showing you, so that shearer here is wearing the technology. So he's out in the catching pen at the moment and so he's wearing a number of different sensors. So you can see there he comes out onto the board. And so this was in the early days. So you might be able to see there's a hunk and MacGyver tape around his mock there. So we don't think that that's probably how we'll do it in the long run, but you know, for proof of concept trial. And you can see there's sensors here on his arms and on his wrists on, he's wearing a belt there around his lower back. So looking at a number of key pressure points. And so what we think is that we are going to now, so that proof of concept was quite successful. We've understood a lot more than we did before about shearer injuries. And so what we're going to look at in the next phase of that is a conceptual design for an active solution. So something that shearer might be able to wear. So whether that might be a belt around their lower back or something to let them know when they're hitting those risk periods. Having a look at the effectiveness of prevention, intervention and fatigue management. So for example, you know, what would be the most effective stretches to do before the start of every run? Or is there a stretch that you could do in the middle of every run that would lower your risk of injury? And also, of course, shed design as well. So trying to straighten up the catch and drag, making sure you've got sloping pens and all those sorts of things that Henry discussed. The other thing that we're looking at at the moment is we did an international expression of interest to have a look at maybe what we could do in terms of improving shearing shed design and around handpieces and that sort of thing. One of the, I think we've got about 35 applications from around the world to that. And the one that we're working with at the moment is to have a look at changing the design of the handpiece so that it would not require a down tube and not require overhead gear and it would be battery operated and that the shearer would wear the battery. What we're also looking at doing is integrating sensors and all that sort of thing into the handpiece so that it would reduce skin cuts and second cuts. And so that maybe it would let you know, if you're coming up off the skin or you're not close up to the skin and whether you're tipping too far forward or back in the angles wrong on the handpiece. It's not sitting on the skin nicely. So what we've been looking at is, and basically I guess what we've used as, what would need to be the absolute minimum would be the battery would need to go for at least two hours so they could complete a run without having to change a battery and that it would at least need to have the power of a Heinegger EVO to run with. And the automated wall handling project is a new project that we've started working on to have a look at whether we can build a semi-autonomous wall handling system to go from the wall table through to bailing. So we're looking at that four main components including automated inspection system to identify fleece contamination, automated skirting and removal of pieces. An optical inspection for automatic wall classification so wall-classing and automated sorting and bailing. Right, now through to wall queue. What is wall queue? Wall queue is an innovative new resource for the Australian wall industry. AWI invested on behalf of Australian wall growers in the wall selling systems review and out of that review overwhelmingly the response was that Australian wall growers were quite keen to see more, I guess more competition for their wall, be able to put their wall into an online system and also to have a bit of a one-stop shop. So out of that came the online platform that is wall queue. So basically, I guess it's a set of digital tools which should allow more informed decision-making and allow you to capture data and get data fed back in from across the supply chain as well. Okay, so an industry network. So the wall queue platform provides an online platform and online message board, all that sort of thing. Growers can get on there, can put their own profile up. So you can put a profile up for your wall growing enterprise, put up some nice pictures of home or put up pictures of shearing, that sort of thing and create your own, I guess sort of like a LinkedIn profile for your farm on the platform. It's also got a built-in e-specie as well. So on some work that we've done, we estimate that over a million dollars could be saved in the wall industry each year if everyone moved to e-species. This is just basically because it would be removing a number of the manual transmission errors and that sort of thing that happened. Sheds are a busy place and sometimes some of the people who are working in the sheds or in the brokerages might not have the best handwriting. So that should just help eliminate some of that human error there. The great thing is as well that the e-specie then stores all of that digitally and keeps it within your wall queue account. And so then what you can get back is the information from your wall testing and this can be automatically uploaded into the back as well. So you see what you sent in based on your e-specie and then you get your wall test results back from that as well and that goes in it. And then that saves year after year after year so that you can keep that as an ongoing record of how your clip has performed in different years. There's also a ready retina. So I'm not sure, some of you here might have used wall check. Over the years, wall check is basically takes in a lot of industry information to understand what prices, what the predicted prices are based on how the market has been selling and ready retina is the updated version of that. And so that is all, it's just an estimation of current market value. And that is all within the wall queue platform as well. And the great thing is you've got your e-specie sitting there, you've got your test results sitting there so you can plug in what your test results say and what you think your estimated price should be. Wall queue market as well. So it'll be the marketplace enabling you to pop your wall on there to brokers to put your wall on for you. And then wall can sit there. So there is also, there's a live auction aspect to it but there's also an opportunity to post wall there for sale. And then that wall can be bought by anyone anytime. They don't have to wait for the auction to be open to do that. And so then again, the flow of information back once your wall is sold, whether it be a traditional auction or you know, on the platform, that information also gets fed back in as well. So you can keep a running tail a year on year of how your wall's performed. In closing, I just want to pop up the AWI R&D contacts there. So the research team is headed up by Jane Littlejohn but all of the contacts there are for you. And then down the bottom, we've also got Stephen Fein and Marius Cumming who are the general managers who work within the consultation wall growth services team which is where Henry works for AWI. So of course, Henry and I are here and more than happy to answer any questions. Of course, you've got John Roberts and Andrew Dennis but we'll queue as well. But thank you very much. And do I have time for questions? Oh yeah, I've got time for questions. Thank you, Emily. Do we have any questions for them? Yes, is the smart tag being developed for virtual fencing? Yeah, so virtual fencing at the moment, the only technology that we know of at the moment which is one that's been co-invested by AWI MLA, Dairy Australia and a few other CSIRO, it's on a collar and it is only available for cattle at this stage. The big problem for sheep with virtual fencing is the wall blocking the sensors, blocking the electrical stimulation. So virtual fencing basically works by, so you would set up an imaginary line in a paddock that you don't want your animals to cross. When the animals approach that line, they will receive an auditory cue so they'll get a loud beep or something and hopefully that'll turn them back. If that doesn't happen, then there's an electrical stimulus and then if the animal proceeds towards that imaginary line, they get another electrical stimulus. So there are some, I guess there are some, few blocks as well to the roll out of virtual fencing. So the company adjacent to have the patent for the virtual fencing for the collars for cattle, there are a number of states which the regulations prohibit any sort of virtual fencing because of the electrical stimulus. So one option that was suggested for sheep was that maybe you could apply the electrical stimulus via an e-tag. In Victoria, that's already banned. And New South Wales currently, you can't use virtual fencing because it's illegal to apply that electrical stimulus that way and then in WA as well, there are some regulatory issues. So there are some issues at the moment to getting that delivered in a broad-scale way. I do know that New South Wales DPI are lobbying the government for a change of regulation at this point. But yeah, I guess it'll just be interesting to see where state regs go. But usually, I mean, in terms of welfare legislation, generally, if it's already banned in one state, it's usually moves forward that way around the other states rather than being repealed. Has any work been done on cameras on sheep? Was that not practical? Cameras? You've got a visual account of what's going on in your flock day by day or what the habits are, daily drinking and working, all that sort of thing. Yeah, so I know we have done some work with predation management, with cameras, and understanding and understanding, you know, using motion sensors and that sort of thing on the perimeters of the paddocks and understanding how the flock's moving and using ID so that we can understand, you know, because obviously a kangaroo could bounce past or whatever and understanding, is it really a dog or is it something else? I mean, in terms of actual understanding behaviours, I know that there has been some used in trial work so that we can understand, you know, if we need to see something specific. But I don't know of any work that's currently underway in terms of impatic behaviour. Yeah. I think there are some water telemetry systems that are currently using cameras. Is that mostly they're just to monitor the water or to monitor the sheik? I think it's both, but not on the sheik themselves. Just with how long do you reckon you still have to work on the tags and stuff and what kind of cost would it be to set up for that kind of thing? Yeah, so a bit of a how long's piece of string at this point. We don't want to release it until it's, you know, robust enough to release. These, at the moment, our aim is to get these to at least be less than $10 a tag at the moment. So we're hoping for somewhere around six bucks, but we're thinking they have to be sub-ten to be a going concern for anybody. Don't tell Caroline or I told you this, but she didn't want me to quote a price, but I think, you know, you've got to tell people. So at the moment, what we're looking at with these, they're not in full scale production or anything, but we'd be hoping that we can get these for 120 per unit or less. And then I guess it would just be from there on, it would be what you want to use the tags for and what functionality you're hoping to get out of them. If you want to do mothering up, every you and lamb would need to have a tag and a collar. If you want to do grazing, you might only need to tag 40% of your mob. For predation, for example, we're getting that work done at the moment to understand how many in the mob we would need tagged. But considering, you know, usually when there's a predation threat, the whole mob will move. We think, you know, you might only maybe need 10% to understand predation, if that's your main concern. But if you want the tags to do everything, the full breadth of what they're capable for, you'd likely need to tag, you tag your whole mob. But then these are battery operated and these are designed to just go onto the lamb. You know, so these only need to be within proximity of the you tag for two days to be 98% accurate. So, you know, you could potentially, you know, obviously I know the labor would be quite a lot, but you could go out and you could collar one mob of yous and lambs and then, you know, you could pull those collars off and then go and collar it, you know, you wouldn't necessarily need to have 2000 collars because you've got 2000 lambs. You can reuse those. Hi, Emily. Emily, is there any work being done with the, to automatic condition scoring of sheep? Yep, I believe there is. Not by us, but there is. I know of two different projects and I've received nothing more than cursory information on those. One of them's been, one of them was, I think one of them's been looked at by University of Adelaide and I believe M1 is also Msaba's DPI. Just for my question, I think South Australia on those states that doesn't currently allow virtual fencing, but the legislation's being looked at and reviewed at the moment, so it's to watch this space. My question's around, I guess, we're being bombarded with data and stuff as it is and we've all got apps and things that we use now. Is there scope for this work to just be integrated into things we're already using, like if we're using AgriWeb or Mayor Grayson or something, that this just puts more into those things rather than another thing that we have to use? 100%. Yes, our aim, so we haven't quite got to that point yet, but yes, our aim would be that that would be able to, that we'd just be able to supply an API and that, so a backdoor data sharing thing to allow that data just to go into some of those other. I think AgriWeb's already receiving WillQ data, isn't it? Is that right? Yeah, yeah, so yeah, we have, I guess, we will certainly build some sort of functionality so that there would be an app or something that you could have if you had these tags so that you could get those alerts directly through, but absolutely we'd be looking to integrate it with other existing platforms, so that in particular things like migrazing and that sort of thing, so that, especially looping grazing stuff and that sort of thing, yeah, 100%. Just a quick one. If DNA testing was a lot cheaper, do we have to worry about any of this collars and EOD tags for use when surely the three put through the DNA mobs if they got a lot more commercial animals through would outweigh the fact they charged a lot of money for that service. Is that all it takes is to drop that DNA price to make the collars and so on superfluous? Yeah, yeah, I reckon it could for sure. I guess the only, I suppose that's why we're sort of looking at the tags, being able to multi-purpose things so that they wouldn't just be doing one thing. So yeah, I think that I think if that's, if people only want the tags for parentage, I guess, yeah, that would be an individual business decision and having a look at, you know, what's the cost outlay for the tags and what's the cost outlay for DNA testing and what's quicker and easier for your business and makes more financial sense, absolutely, yeah.