To be honest, it’s hard to tell.
I think that automated spraying by swarm bots, drones and full-size SP rigs will become very common in the very near future.
But which one is going to win?
Will it be a big SP sprayer that is driverless, or should we be thinking of having multiple, small machines?
We’ve heard some great stories about the swarm bot towing a 24m boom spray with WeedIT sensors for summer spraying and winter fallow. Weedsmart interviewed Cameron Jenkin who works for the Flannagan brothers in the north-eastern wheatbelt of WA. Their swarm bot, “Alfie”, turns up for work every day, gets a fresh tank of fuel and chemical, and heads out spot spraying for 24 hours at a time. Alfie has been so successful the Flannagans have ordered a second machine. You can hear the interview here https://www.weedsmart.org.au/audio/regional-update-cameron-jenkin-farm-manager-pindar-wa/
It seems like these things are going to be a no brainer for growers with large areas of summer spraying and winter fallow, but what about everyone else?
Facts and figures
Swarm bots are built by Swarm Farm in Queensland. They are an automated platform powered by a 83-horsepower diesel engine over hydraulic drive. They travel at about 10kph and can tow a 24m boom spray with a 3000L tank. Swarm Farm were originally running a lease only plan, but now you can buy one for about $305K. A 24m Hayes boom spray with WeedIT sensors will cost you about $330K. There are several boom sprays that can be teamed up with the swarm bot.
The machine can fold itself up and drive back to the tank and wait for someone to come and fill it up. For low weed densities it can run for two days on a tank of spray, but it needs fuel after 24 hours. In big, open paddocks, they can cover about 450ha in a 24-hour period of spot spraying.
I’m excited about the low cost of maintenance. The first three years of a swarm bot is under warranty, so it’s just oil changes, and given it’s a small engine it’s not even a lot of oil. A new engine is only $13K and a new wheel motor is $4K.
I think a lot of growers are also excited about not getting up at 2am to go spraying summer weeds!
Jobs for humans
Some people may say that we are doing humans out of a job. And we are. We are doing them out of a long, boring job that is currently difficult to staff. In the words of Queensland farmer, Jamie Grant, who we also interviewed for the Weedsmart podcast, “My workers all still have a job, but now they are doing more rewarding work that they like, rather than long hours on the boomspray”. He also commented that the swarm bot totally changed his farming system. He now has the machines out there spraying more regularly, targeting much smaller weeds.
Swarm bot alone?
It’s ambitious now to think that the average 4000ha farm would get away with only a swarm bot and nothing else, but this is what the company is shooting for.
I set about trying to do some numbers, here’s a summary of what I came up with.
** SP = self propelled boomspray. GoB = green on brown cameras
*** I have used an average of five years of depreciation at 10% rather than just a single year depreciation of the new machine as we assume each machine will be kept for at least 5 years.
Firstly, the ownership cost of the swarm bot is about half that of a new SP with cameras because it is about half the capital cost.
But the SP can do more hectares in an hour/day.
The SP gets the spraying done in 420 hours, but it takes the swarm bot 1738 hours to do the same area of spraying. And this table is flawed. I have assumed the swarm bot can spray for 23 hours a day, which may be true for spot spraying in summer, but will not be true for winter bulk spraying.
Even though the swarm bot only uses 5 to 8 litres of fuel per hour, it ends up using more fuel per year than the SP because it covers less hectares per hour. And yes, I guessed at the repairs cost.
When we add it all up, including labour and subscription cost for the swarm bot ($12K) it appears that the swarm bot is cheaper than the SP, and bonus – you don’t have to drive it!
But….
And it’s a big but. This comparison isn’t fair. The swarm bot will simply not be able to cover the in-crop spraying in a timely fashion because it’s slow and we know that we have limited spray days.
Two Swarm Bots
The question is, could a farm get away with just two swarm bots and no SP? The total annual cost of two Swarm Bots comes in at $268K, which is very similar to one SP.
One Swarm Bot plus one old tug along?
Or could the average farm get away with just one swarm bot and a cheap tug along that gets pulled out for a few weeks of the year to get on top of spraying when you need to smash hectares? To be honest, I don’t know. The numbers say yes, but are the logistics achievable?
For big farms, the swarm bot may just do summer spraying, winter fallow spot spraying, and perhaps a bit of bulk crop spraying with the aim of reducing depreciation/hours/maintenance of the main SP. This seems very plausible, and several farms are already going this way.
Next steps for Swarm Farm
Autofill is coming next year where the bot pulls up to a nurse tank and is filled automatically. It will be a fixed point at first, then they’ll work on a nurse truck in the future. This will be additional cost, but this could be the thing that makes bulk spraying feasible. The grower would need to set up a nurse tank with a brew in the morning, then leave the bot to it for the rest of the day.
Swarm Farm is now starting to look at diesel over electric (currently diesel over hydraulic). This is likely to reduce fuel and maintenance costs, and possibly speed it up a bit as well. If this speeds them up by two or three kilometres per hour, this will make a huge difference. But it is likely to be a couple of years away.
Swarm bot will be able to do other jobs. Wheel track renovation and mouse/slug baiting has been spoken about. I’m sure there’s other jobs as well.
Other machines similar to Swarm Bot
Another one to check out from the Netherlands is the AGXEED AgBot. This is a diesel over electric bot with a 156-horsepower engine. Beefwood Farms in NSW have purchased five of them. I know that they are towing a 36m boom spray with WeedIT sensors as well as doing wheel track renovation, but I’m not sure what other jobs they have put it to. https://www.agxeed.com/
When will an SP be driverless?
I asked this question of Anton Kowalenko from John Deere at a recent Weedsmart event. His answer was 2030 because the CEO of John Deere had said words to this effect at a shareholder meeting recently. Once we know the cost and the speed etc., we will be able to do our numbers to work out whether we should be thinking of multiple Swarm Bot style machines versus driverless versions of our current machines. The likely answer is that both options will be out there in number soon.
Summary
Could the future be a swarm bot or two and a couple of drones for all your spraying needs? If I said that a couple of years ago you would have said I was totally crazy. Now it seems almost plausible. I’m going to bet that within a couple of years we see farmers doing it. Autonomy will be part of the future of Australian agriculture, the big question is whether it will be large or small machines. Now, the cost of the large machines makes me think that the small machines will win, but I could be wrong.
It blows me away that some young French guys with no ag experience have beaten the big companies at ‘see and spray’ technology in crop (Bilberry), and now an Australian farmer has beaten them to autonomous Agriculture (Swarm Farm). We have the innovators, now we just need to put these innovations to work in the most profitable way.
