Compared to more predictable environments where warehouse robotics have flourished in recent years, field robotics innovators must contend with the vagaries of weather, varying terrain and perishable crops that are sensitive to damage.
The two sub-sectors are like apples and oranges you could say, which are exactly what Ripe Robotics is harvesting with its automated prototype Eve, a biblical reference in this emerging sector as co-founders Leopold Lucas and Hunter Jay chase the "forbidden fruit" with their startup.
The problem these entrepreneurs seek to solve is the effects of labour shortages that weigh heavily on the fruit industry. The Australian Government announced a new agricultural visa to address the issue yesterday, but the challenge is unlikely to go away anytime soon.
And it is not just limited to Australia. In May, one of the USA's leading fresh produce industry groups Western Growers launched its inaugural Global Harvest Automation Initiative cohort to address this very topic.
Sydney-founded, Shepparton-based Ripe Robotics was one of just 13 companies worldwide selected for this prestigious, collaborative group, and more recently was picked as part of accelerator Startmate's winter cohort of Australian and New Zealand startups.
At Business News Australia, we catch up with Leopold Lucas who discusses the company's vision to tackle an addressable market of "a trillion pieces of fruit".
From its origins with the University of Sydney's INCUBATE program in 2019, Ripe Robotics is now fielding calls from all over the world asking when a commercial fruit-picking robot will available, but they will have to wait as Australian apple, citrus and stonefruit growers test out the developing technology.
Lucas says the project grew from Jay's interest in pushing the boundaries of what artificial intelligence can achieve, combined with their common desire to "produce something that's meaningful and valuable".
"It ended up coming down to a few different areas in Australia, particularly mining and agriculture, but we realised that agriculture has a much bigger opportunity for expansion and has really been left behind by innovation that's benefited a lot of other sectors," he says.
"We saw that if we could solve this problem, there was a huge opportunity to benefit a lot of Australian growers and growers all around the world who are continually struggling with labour shortage problems, and also the waste of fruit that gets left on the tree or rots every single year."
First there was the idea, but then came the practical realities of dealing with an industry neither had experience in. Involvement with INCUBATE was critical for guiding the two engineers and entrepreneurs-to-be to understand the market opportunity.
"We ended up calling farmers all over Australia. We visited a bunch of different farms, and the Department of Primary Industries were really open to us and actually arranged a trip for us to go out to Batlow and Orange and meet a lot of different growers," Lucas says.
"That really confirmed our hypothesis that this was a major problem, that people really did want a solution, and then if we could build this robot there would be a significant amount of demand."
He says apples were chosen as the first target crop due to the way they are now produced on two-dimensional trellises in many cases, enabling growers to get better yields and fruit quality.
"This has the added benefit of being really great for robotics, because it meant that you have larger spacing between the rows and you have more visibility of more of the fruit, so it’s easier for a robot to pick and navigate the orchard autonomously," Lucas explains.
"The other benefit and the reason for apples is that there is a really high cost of damage. If a human damages on average 10-20 per cent of the fruit that they pick, if you can reduce that you've got a massive benefit."
But there were of course technological challenges to be overcome, and this will likely be ongoing.
"We had to be able to develop an AI system to identify the fruit, to identify if it was ripe or not, and to work out where it was in the tree, and then to identify the best sequence of fruit to pick and also to avoid obstacles that it might run into," Lucas says.
"The next thing was the hardware, and this was much more complex because we didn’t have a lot of experience in hardware; it was a big learning curve.
"Hunter basically led that and developed designs, and we built an initial lab prototype which was just literally built out of a Bunnings bookshelf."
This gave the pair the foundations they needed to raise a pre-seed round of just under $100,000, which enabled Ripe Robotics to build its second version prototype called Clive, optimised for the field so it could be used in trials.
"We did our first trials in Bilpin in early 2020, and then around the Hawkesbury region around Richmond and then down in Griffith on oranges, and that gave us a whole lot of other learnings," he says.
"It's very important with field robotics to actually get out there and test it – you can’t sit in your house and build these robots and assume they’re going to work because the farm environment is so unstructured and complex.
"There's so many extremities you've got to factor in for, so obviously being able to pick at night-time, being able to pick in different environments – if there’s rain for example, being able to navigate the ground and not tip over and avoid obstacles that could be on the ground, and long run times."
The entrepreneurs have also found it is actually ideal for the robot to pick the fruit at night because the lighting is more consistent.
"During the day you have to account for whether it’s a morning light or an afternoon light, or different spectrums, different colours to consider, and if it’s cloudy or it’s raining, there’s so many more types of data you collect," he says.
"Whereas if it’s night-time and you have a consistent light source then it’s actually easier for the systems to accurately detect the fruit."
Lucas explains this range of hurdles to overcome with field robotics are part of its appeal.
"Field robotics is the cutting edge of robotics at the moment, which is also another reason we looked at doing it. There’s been huge amounts of innovation in warehouse robotics, and all of those things, and that’s because they're structural environments that are quite predictable," he says.
The Ripe Robotics co-founder admits Clive would break down every now and then because it wasn't upgraded to be commercial. However, with some key tweaks the company has rolled out the third-generation robot Eve.
Having just closed a seed round that is worth approximately $610,000, Lucas explains Eve is a fully commercial machine "in its infancy".
"We've got everything we need on this machine to make it commercial to be able to actually carry fruit in bins and operate autonomously," Lucas says.
"There are a huge amount of upgrades from our last one, but really it's about simplifying the amount of parts that we’re using, building it to be more robust and able to deal with the farm environment and avoid bumps and all the tricky things of actually moving around, and to be able to reliably pick the fruit."
This winter the robot has started trials with Mitchell McNab, a prominent Shepparton fruit grower and the chairperson of Fruit Growers Victoria. Ripe Robotics is also in discussions for a potential trial with one of Australia's leading orange growers, while Eve will be used to pick stonefruit for the first time later this year when the summer fruit season begins.
"The name is probably inspired from some popular culture or religious references. We sort of see it as the picker of the forbidden fruit," Lucas jokes.
"We saw that it was important to work on multiple fruit types, and that’s because we’re going to provide our system as a service and we want to make sure we can actually use it year-round.
"It wasn't necessarily our intention in the beginning to do all of these fruits, but we realised that the way we designed it actually did work on different fruit types without too much change."
The challenge with stonefruit will be the short harvest window to get the process right, but this could also be a positive for Ripe Robotics.
"Humans struggle to find the exact right time when it's really good to pick stonefruit, the optimal time, and so there can be benefits if you have an automated system that can actually precisely detect how ripe it is," the entrepreneur says.
"They also struggle often to get enough people because there’s a very short window where they need a lot of people, and there’s always a lot of different crops that are competing for people and obviously the cost is huge if they don’t manage to get to their fruit."
Lucas says the pre-seed funding got the company to where it is today with a very resourceful approach, but he is happy to be involved with Startmate as Ripe Robotics works towards the next steps in its growth journey.
"We really wanted to work with them, they’re the most reputable accelerator in the country, they’ve had an incredible success rate with a lot of startups that have gone through them," he says.
"The benefit to us was there’s a financial component, but the main thing was getting access to expertise around fundraising, around hiring, around deep technology, around startups, and having a community of really invested mentors."
He says companies in the ag-tech space need the support of investors who know it’s not going to be a quick win.
"It's not like next year we just flip the company and we make heaps and heaps of money – agtech is more I’d say patient capital in some sense; patient but smart capital where people realise that this is a huge market, there's a massive opportunity, there are big competitive barriers," he says.
"There’s a lot of innovation that goes into this and there’s a lot of value that can be created. It’s a different model, and I think the challenge is, we talk to probably close to 100 investors, and had a lot of challenges because the industry just really isn’t set up to fund hardware in Australia.
"A lot of funds are really focused on software, and Software as a Service (SaaS), but I think that's starting to change. There's definitely some really great funds that are that are specialising more in hardware and looking into those areas."
Lucas doesn't expect to expand into many other crops in the next three to five years, but in that time he does expect Ripe Robotics to go global.
"At the end of the day, if you just count apples, oranges and stonefruit, there's well over a trillion pieces of fruit picked a year, so to get up to that scale where we’re picking a substantial amount of it autonomously is going to take some time," he says.
"We're based in Shepparton which is the fruit bowl of Australia at the moment, and we’re looking to start to service as many farms as we can in that region, and we’ll be looking for opportunities to expand globally - there's a number of regions in the US and New Zealand that we've identified that we'd like to try and support.
"We’ve had growers reach out to us from South Africa, Brazil, the US, New Zealand, Europe, asking 'Can we buy your robot?' or 'When is it available?' So we know there’s a huge pull globally, but we want to focus on Australia in the beginning to prove out the technology and the operational model, and then be able to expand."
Help us deliver quality journalism to you.
As a free and independent news site providing daily updates
during a period of unprecedented challenges for businesses everywhere
we call on your support