Long Read: Robotics, Machine Learning and Food Security – An interview with Dr Duncan Robertson from Dogtooth Technologies

As sponsors of the Cambridge Independent Science and Technology Awards AgriTech of the Year Award, we’ve been keeping a close eye on the finalists.  As we prepared for the Awards ceremony on the 1st November 2018, we showcased a bit more about the four finalists as exciting members of the Cambridge AgriTech cluster.

We caught up with Dr Duncan Robertson, CEO and co-founder of Dogtooth Technologies who won the prize and were crowned ‘AgriTech Company of the Year’.

All about Dogtooth Technologies

Over our long, hot summer in the UK, there were many headlines about fruit being left to rot in the fields due to a lack of pickers – and these trends are set to continue. Melbourn-based start-up Dogtooth Technologies seeks to address this problem using state-of-the-art autonomous fruit picking robots. They are currently manufacturing and testing the third generation of their strawberry picking robot. It can navigate within polytunnels, find ripe berries, pick them, perform a visual inspection, and place those suitable for sale into punnets. Dogtooth’s offering combines robotics, machine learning and AI, and a robotic workforce that can move around the world – much as human pickers do.

Duncan has strong connections with Cambridge, having obtained a PhD in Machine Learning and Computer Vision from Cambridge University Engineering Department. He was also co-founder of Cambridge start-up Metail.

Why AgriTech?

We asked Duncan why he decided to apply his skills to AgriTech, rather than another sector.

On the face of it there are several problems we could have tackled. However, we were interested in exploiting machine learning in a more tangible way than many of the start-ups we know – we didn’t want to develop yet another mobile phone app. Ed and I were also very interested in agriculture because of increasing concerns about food security and the sustainability of food production.

Having made the decision to pursue AgriTech, specifically robotics and machine learning, Duncan and co-founder Ed Herbert decided to develop their first strawberry picking robot for two reasons:

Firstly robotic soft fruit picking represented the cutting edge – it was a problem difficult enough that no one had solved it yet, but not so difficult that it wasn’t amenable to solution by people with our skills and IP. Secondly strawberries are a great British product and here in Cambridge we have many growers on our doorstep.

Scale-up Challenges and Opportunities

Since Dogtooth are in the process of scaling up their manufacturing and operational capabilities, we asked Duncan what he thought were the biggest challenges and opportunities facing AgriTech companies seeking to scale globally.

It is clear that the labour problems facing growers face are not peculiarly British. Growers around the world report difficulty in recruiting picking workers and this seems to be a consequence of global phenomena – such as changing attitudes of young people towards agricultural work and the availability of alternative forms of employment.  But these changes have created the opportunity that we seek to exploit.  We don’t seek to displace people, but as an industry we do need urgently to fill the recruitment gap.

Duncan also sees capital utilisation as a key opportunity for the industry.

 Harvest seasons are short – but there is no reason why robot capital can’t migrate the way humans did to follow the growing season.

As SmithsonHill has highlighted before, Duncan agrees that Agritech is key to sustainability progress.

Modern farming methods including robotic harvesting can help to make our farming more resource-efficient, minimise water consumption and reduce dependency on chemicals. Robotics has a significant role to play in improving sustainability in the sector.

Productivity

In the UK, the government’s Industrial Strategy highlights AgriTech as a key way to meet productivity targets in the future.  We asked Duncan about how his technology helps improve productivity.

We tend to think of productivity in terms of fruit harvested per pound spent. On this basis, the cost of our robotic offering will certainly be competitive with the cost of human picking labour. But a key benefit of robotic picking is to give growers predictable harvesting capacity, thereby reducing waste. By contrast, recruiting human workers is nearly impossible at short notice. Our aim, however, is not to take jobs away from people. If growers can’t address their labour problems soon, then they will grow less fruit or go out of business – eliminating jobs. By providing growers with the tools they need to make more effective of the human labour that is available, we hope that our robotic picking solution will enable the industry to thrive – thereby creating jobs in the areas of horticulture that are not automated.

Global Food Security

At SmithsonHill, improving global food security is a key driver for us – we asked Duncan what he thought is the single most important thing to improve global food security.

When you look at the numbers, demonstrably one of the most important aspects is reducing food waste.  A massive proportion of the food we grow is wasted before it even reaches the consumer – up to around 30% for some crops.  This is something we need to tackle urgently.  Robotic picking helps by improving the efficiency of whole supply chain.

Agricultural Revolution

At SmithsonHill, we see AgriTech as the beginning of the next agricultural revolution – we asked Duncan if he thought one was needed.

Revolutions always proceed one small step at a time.  Good engineering is about many small changes and few big ones. We need to drive significant change, but this will happen by degree, which is the nature of engineering.

Duncan added,

We are in a period of unprecedented progress in machine learning. We’re seeing the possibility of using computers to solve problems that would have been considered impossible 10 years ago. In particular, machine learning has allowed computers to obtain a much more accurate interpretations of the world around them through images and to make much more effective use of resources in achieving complex goals. It is these advances on which our intelligent low-cost robots are based. We are standing on the shoulders of giants, leveraging the progress in AI to tackle problems we couldn’t before.

You can read more about Dogtooth and their work in this article by Cambridge Independent.  We will definitely be watching Dogtooth’s progress and hope to see robots in a field near us soon.