interview

Over the years, the typical number of cows on a farm has steadily increased.

More cows inevitably mean the farmer has less time per cow to check on fertility, health and general wellbeing. Which increases the risk they miss the first sign of an emerging issue. IoT holds promises for a solution. Sensors never get tired of watching the herd 24/7. Algorithms for pattern recognition detect even the slightest change in behaviour. And last but not least, machine learning techniques help to identify the issue that may have caused the detected behaviour change. Even before the issue becomes visually apparent. Thus, IoT technology helps farmers to focus their attention on cows that really need it.

Within the Dairy Trial, two use cases work on an application simultaneously.

Within Use Case 2.2 Happy Cow, start-up Connecterra has developed and tested their IDA (Intelligent Dairy Farmer’s Assistant) application and has started selling their product internationally. In Use Case 2.3 Herdsman+, the University of Strathclyde is investigating how sensor information can be enriched with multi-source data input and how it is translated into relevant information to support the farmer. Both Use Cases share their progress up to date, the challenges met along the way and the next steps they intend to take. Robyn Lee Bonnin, marketing manager at Connecterra is one of the people working on Use Case 2.2 Happy Cow. Ivan Andonovic, from the University of Strathclyde, is the principal investigator in Use Case 2.3 Herdsman+.

Challenges

‘Herdsman+ is primarily intended for housed cattle, which is common practice in the UK due to the weather conditions,’ Andonovic explains. ‘Our research builds on a system that we helped develop the technology for. It’s a collar-based system, the output of which can be enhanced through combining inputs not only from the cow, but also from the milk robot and the feeding station. This requires data integration between different systems, which is a major challenge to solve.’ Connecterra’s IDA app relies on location and movement data from cows. Bonnin: ‘Our aim is to make the life of farmers easier, wherever they are situated and regardless of their farming system. Now that we gain a lot of interest from farmers, our biggest challenges have to do with scaling and international expansion. Which is not much to complain about.’

Data value

The question of data ownership and the value of data is a constant discussion within all IoF2020 Use Cases. Andonovic: ‘We try to monetize the sharing of data. Data in itself is only the raw material. The real value is in the combination of the data from multiple sources over time. Only then the data starts to tell a story, which ultimately leads to more profitable herd management. We try to show this in our use case, so that farmers overcome their reticence to use systems like this. Farmers who share their data, should always reap the benefit.’ Bonnin: ‘We totally agree on that. It’s in our own interest to create a product that helps the farmer to save money and time. IDA is worth the monthly subscription only when the farmer immediately sees a ROI, a gain in efficiency and farm optimisation. We expect to help reduce farm overhead costs by 20%.’

‘Data in itself is only raw material. The real value is in the combination of data from multiple sources over time.’ Ivan Andonovic - UC 2.3 Herdsman+

Being part of IoF2020

‘IoF2020 provides a great knowledge base for us. Being part of the project, allows us to reflect on what we do, which makes our business more solid. When we have a question, we can always contact one of the partners,’ says Bonnin. Andonovic agrees, but also comments: ‘Up till now, we’ve concentrated a lot on our own research, the creation of a minimum viable product and on installing the system on three different farms in the UK. We do discuss things within the trial, but day to day work tends to get in the way sometimes.’

The farm of the future

IThe accuracy with which both applications predict an emerging issue before they become visually apparent is striking. So, does the farmer of the future still need to know anything at all about cow health? Bonnin: ‘Of course! The responsibility for the cows and the farm always stays with the farmer. The farmer takes the action, that doesn’t change. You should think of IDA as an assistant or an apprentice. Based on data and patterns that IDA understands to be normal for a specific cow, group or herd, the application offers the farmer an explanation of a possible cause, as well as a recommendation for alleviating the issue. In order for the application to learn and become quicker, it is important that the farmer validates and interacts with IDA. The farmer’s input helps the system to improve over time, just like a real-life assistant would.’

Why did IoF2020 admit two similar use cases in the same trail?

‘One of the IoF2020 goals is to overcome the obstacles that currently hamper the uptake of IoT technology. Interoperability is one of those issues. We are interested in what needs to be done to make IoT systems truly interoperable. By admitting two ‘similar’ use cases with a different starting point, one with a research focus and one with a business focus, we hope to get a deeper insight that will help advance the use

of IoT in agriculture.’ - Kees Lokhorst, Ecosystem Chair of the Dairy Trial.

UC 2.2 Happy Cow in short:

The minimum viable product is installed at three test farms in the Netherlands, Belgium

and Germany.

Partners use cases:

UC 2.3 Herdsman+ in short:

The minimum viable product is installed at three test farms in the UK.

Partners use cases:

Ivan Andonovic &

Robyn Lee Bonnin,

Ivan Andonovic, professor at the University of Strathclyde & Robyn Lee Bonnin, marketing manager at Connecterra

Author: Renske Solkesz, Schuttelaar & Partners

use case

Happy

Cow and

Herdsman+