 Hi, I'm Lucy Reid, Ag Tech Extension Officer at South Australian Department of Primary Industries and Regions. We're here at a property in Port Arthur in the Northern York Peninsula, and we're here to see the Rubicon 9000 sprayer for hardy with the GeoSelect technology. Currently we're here with Andrew Snowball and Heath Thompson from hardy. So for all the farmers and that out there that love things like horsepower, how many horses are we talking under that hood? So the Rubicon has got 375 horsepower, we've got a road speed of up to 50km an hour. It's a self-propelled unit, so combined all in one tank boom the whole lot. There are options on boom sizes from 36 metres to 54 metres, which I believe is the largest, with different nozzle technologies that go onto that. So this is assembled in Adelaide, South Australia, and the GeoSelect part is obviously designed and made also in Adelaide, South Australia, but for Australian conditions. So the only additional parts over and above a normal sprayer are the four GPS units. So you've got two in the centre and then one on either boom. That measures the angle of the boom, as opposed to just assuming it's a rigid structure. And then this one here has got 49 individual sections of a metre wide, that it's got airlines and a 12 volt electric system, and the associated wiring and stuff for that. But that is the only additional parts, there's no extra can systems, no extra cameras, no extra power systems. I guess one of the things we wanted to do again is not put so much hardware on the boom. So it's easier for maintenance and fixing and... Yeah, I guess it comes back to that design where we want the farmer to be able to fix the machine if there is a problem, fix the machine themselves or have the best chance to do that. Hopefully with your boxer tools and stuff like that you can make a few adjustments and get going again. Each section of a metre wide is actuated pneumatically or by air, which is done through a solenoid, which is a 12 volt solenoid. And can I just ask why the four GPS receivers? Is it for a good number or do you need more or less? So there's two in the middle centre section to get a very accurate heading and then two either side to know where the boom wings are. So you actually see in the user interface as we swing around, which we will, that you actually see the boom swing around. The boom is not actually one rigid part, it actually flexes. And in bigger booms it flexes even more, so if I was to push that back there, actually at the end there it's a couple metres. So that can actually be the difference between hitting and missing a selective spraying operation. Some camera sprayers, whatever is in front of it, if you put wave a bit of grass in front of it, it will spray. This knows where it has and hasn't sprayed, so if you are turning and one of the booms moves backwards over that area, it won't spray that area. So Andrew, just talk us through GeoSelect. What's the process from start to finish with this? In summer, the first thing that the farmer would do is get a drone operator to come out and fly the paddock. So that would capture an image of the paddock. That's all then processed and put into a dinomap, which is basically a weed map of your paddock. So everything is there prior to actually moving the sprayer into the paddock. So all the decisions about how much to use, whether you want to actually spray the paddock, how fast you want to go, which way you want to run your AB lines, it can already be determined before you get in the paddock. Once that you've got that data, it's either transferred by the cloud or a USB stick or something similar into the sprayer and then it's as easy as just rolling up, turning it on and driving around. Andrew is about to start the Rubicon up. OK, so I'll start the machine, fold it up and then give you a yell and then I'll unfold it. The sense that we have can pick weeds up down to about five cent piece. And what we realised that when we started using this high resolution data that having a single GPS antenna on the roof of a sprayer was not sufficient, accuracy wise. To work with the high resolution data that we had, so you'll see on the boom as it passes through that you've got four GPS antennas on the boom. And what that's doing is measuring where the nozzles are down to a centimetre and allowing the scan data to be most effective in the sprayer. It's particularly useful to have selective spraying on large booms because of the amount of air you can cover. You save on fuel costs and amount of hours on the paddock over a 36 for example. And you're able to get a lot more done in a lot shorter period of time. So the moment we're going to spray this paddock with 100 litres of hectare and the geo-select will work out where that 100 litres hectare has got to go and where it doesn't have to go. So there's various views on the Rubicon, but we'll get going there. So we get rolling. OK, so just turn a bit spraying to get our boom height going. As you can see the boom switch on and off as well as the boom yaw. So you'll see the booms move on the screen as well as out in the paddock. So at the moment we're doing only 10km an hour, but we'll speed it up from there. So we're using Norak auto height on the boom. It's important with a large boom like that to have some auto-leveling capability. The Norak system uses ultrasonic and we've used it and to work quite closely with Topcon to implement the system on our large booms. The nozzles are going off using the drone data. I don't know if you can see in the camera, but we've got LED lights to turn on when the nozzles activate and we do that. In the evening when the light starts to come down a bit, it makes it a lot easier to see what's spraying and what's not. On this particular Rubicon we're running two nozzles per section at a 500mm spacing. So you've got one metre per section going off at any one time. Because the nozzles are not even fan, the software is configured in such a way that you're never half spraying a weed. It's allowing for the uneven nature of a traditional nozzle. On this system we're running two separate spray algorithms. You've got a H-Select system which is able to blanket spray at a varying rate and you've got a selective spray system which is a single nozzle out of the four nozzle system. The system knows where it is by the GPS units on the boom. So when you feed a map into it, it is basically as you move around the paddock, it knows there's a weed coming up, say 500 metres down there, there's one single weed. It knows that there is a weed coming up and whatever part of the boom passes over that weed, it will command the unit to spray that particular section at that point. Whether you're turning, whether you're reversing, whether you're going 25kmh or 3kmh, it doesn't matter. When you pass over that weed, it knows it's there and it will turn on the appropriate section. The other thing it lets you do is in heavier areas, it can recommend that there's a heavy area coming up for weeds to slow down and you'll save money. As opposed to you can have a real light section or even nothing for a long time and it can just say, do 35 until you get to the next area because you're not actually going to spray. Would you say that this was a good system for time efficiency? So it's a little bit of a change in the workflow. I guess it's not just jumping the sprayer and out to go. But I guess most of the times that we've talked to people that in summer spraying day, farmers know they're going to be spraying well before they actually are going to be spraying. In that time that you would have the drone operator come out and scan and then if you're scanning today, say the paddock today, scanning today, then it'd be spraying tomorrow. So with being able to image exactly where your weeds are, so you are spot spraying, what would you say the cost efficiency would be or savings? So if you've got all the weeds pushed it up into the corner, it'd only be one or two percent of the area that you actually need to spray. Obviously you're going to spray more than that one or two percent because you're going to have overlaps here and there. I mean what we say is up to 90 but depends on what your paddocks like and how many actual weeds you have in there and also how fast you run. The faster you run over that map, the more you'll spray. The nozzles can only turn on and off in a certain amount of time. So if there isn't enough time for the nozzle to turn off before the next weed comes up, it will just stay on. So you're definitely getting more paddocks per tank than what you normally would. Less time having to go back to the farm, fill up, you know, put more chemical in and stuff. Plus strategic decision making is that you actually know what's out there before you get there. So you can plan logistically for chemical deliveries or whatever else is going on rather than, you know, at the moment I'm just buying how many hectares I've got. At the moment it's only fallow weeds that we're picking up but into the future potentially if you can detect it with, you know, a drone or a sensor or something like that then if you can detect fungal and you can map it, you'll be able to do it. I guess that leads on to the old green-on-green question, you know, same deal. When sensors and that are able to pick up reliably a green-on-green so you've got, I don't know, ryegrass in a wheat paddock or something else in lentils and, you know, turnips or whatever you've got, if that's able to be mapped then this is able to be done. So there's no change in hardware on the machine for doing other applications as long as it can be detected by the imagery done at the start. For this rig, who would your ideal customer or target market be? For the Rubicon because of the size and the boom width it'd be your large broad-acre farmers, contractors, that sort of thing. Thank you, Hardie, for inviting us out to see the Rubicon 9000. Thank you to Paul Jarrett for letting us see your property and utilise it for today. If you'd like to find out more about ag-based technology in South Australia please head to the Department of Primary Industries and Regions website.