I spent Wednesday at Borgeby Faältdagar, the professional agricultural fair outside of Bjärred in southern Sweden. I had heard about the fair from the people at SenseFarm who had attended the AgrIoT meeting at ITU on Tuesday night. The idea was to get a better hands-on sense of what modern farm equipment looks and feels like, to talk to people throughout the industry (farmers, suppliers, manufacturers) about the ideas and trends in the field, and to perhaps make some research connections. I managed to accomplish most of that, while also shooting some decent photos and video, and recording a few impromptu interviews.
One of the first things I noticed at the fair was the size of the equipment that was on display, and especially how the larger machines were featured and advertised. While there was a range of equipment from remote sensor systems to animal enclosures, it was clear that the tractors were the main draw. There were regular demonstrations on large sample plots, virtual reality tractor simulations, and row after row of massive vehicles.
I spoke to a few representatives about the size of the machinery and two in particular seemed to have spent time thinking about the advantages of smaller equipment. The first, a rep from Yara (a company that sells sensing systems and fertilizers), mentioned that smaller machines would scale nicely and one could use the same equipment for both small and large farms. He also mentioned the versatility of being able to switch some of these smaller machines over to a different function to split tasks, as well as the minimization of soil compaction given smaller machinery. Finally, he pointed out that there’s inherent safety factors in automating smaller equipment versus large tractors. A rep from Fendt had similar things to say about their Project MARS research.
One thing that seemed to come up quite a bit from the representatives is that they were unsure of whether farmers themselves were interested in smaller machinery. A few opined that the farmers sought large equipment in part for its largesse and the present advertising seemed to indicate a similar set of assumptions. At Laura’s suggestion, these observations may eventually end up in a paper about the aesthetics of farm equipment.
One of the things that surprised me most was the response of the union representative that I spoke to. He belonged to Kommuna, the Swedish Municipal Workers’ Union and I asked him about the union’s view of robotics in farming.
From my experience, the “conventional” viewpoint is that robotics and automation pose a significant challenge for workers (and therefore unions) as concerns arise that they will displace certain jobs (see here and here). It is not that unions oppose the use of technology per se, but there is a very real concern that jobs – especially those that support the most financially vulnerable of us – will be displace by automation.
However, the representative from Kommuna seemed to take a different tack. His view was that workers that once needed to do jobs that occurred at odd hours and caused repetitive strain injuries (milking cows was his example) could now be transformed into more supervisory roles that could mean more of a regular 9-to-5 day. Though he did not explain (or perhaps I did not understand) how this would not incur the loss of jobs (machine supervision typically requires less workers than manual labour), I suspect that this might result from the union’s broad representation.
Kommunal represents people who “take care of the elderly, do lunches for school pupils, and take care of children in preschool…drive buses, put out fires, sweep chimneys and drive ambulances…look after the animals kept in zoos, tend to the green on the golf course and drive combine harvesters in the country’s fields.” That is a pretty broad base of workers and I suspect that Kommunal might have an eye towards retraining displaced workers for other roles in their diverse areas of practice.
Of course, one of the main purposes of the visit was to assess the state of the art in relation to my own project. Much of the most closely related work was in the realm of sensors, most probably because this represents the lowest-hanging fruit. There are a number of efforts to place permanent sensory systems in the field (for example, the good folks at SenseFarm, who first told me about Borgeby Fältdagar). Other efforts, such as the Yara N-Sensor place sensors on movable platforms. That particular one is mounted on a tractor and attempts to measure nitrogen uptake in plants using camera vision, during the fertilizer application process. Finally, another approach is that of the Solvi and other similar projects. Solvi is actually an vision-based web app that allows farmers to map a field with an off-the-shelf drone, and then assesses nitrogen uptake (in a similar manner to the N-Sensor) from those photos.
These efforts are all interesting, but none of them represent an effort to introduce autonomous robotics into the agricultural field (never mind exploring non-conventional agricultural practices with said robots). In fact, I was told by one representative at the exhibition that Sweden is actually a rather conservative market for agricultural equipment and that typically technologies will have been field-tested for 5 years before they hit the Swedish market.
I did hear about a project, however, while I was at the fair and took the opportunity to talk to a rep from the company responsible for it. MARS, or Mobile Agricultural Robot Swarms, is an EU-funded project run by Fendt (makers of the 18 tonne combine) and The University of Applied Sciences Ulm. MARS purports to have created a fleet of medium-sized (about 1 cubic metre, maybe a little larger) robots that perform agricultural tasks. If one detects a hint of skepticism, it is the result of a lack of photographic and video content that shows live-action demonstrations of the robots (as opposed to animated renderings). The lone paper published from the project, “Motion Control for Omni-Drive Servicerobots Under Kinematic Dynamic and Shape Constraints”, is so generally titled that it could apply to nearly any robotics project, and mentions the MARS project only in the funding acknowledgements. It may well be quite successful, however, so I’ll have to see if I can get an invite to one of their test plots.
Even taking the project at face-value, MARS can be seen as a effectively implementing the same industrial system of agriculture that exists today, only with smaller machinery. It’s an interesting first step towards what I am proposing, but it does not really attempt to change the way agriculture functions to suit the new machinery.
I spoke to some researchers with the SITES project from SLU (Sveriges Lantbruksuniversitet, the Swedish University of Agricultural Sciences). SITES is Swedish Infrastructure for Ecosystem Science and they have a series of field sites around the country. The site in Lönnstorp is near Malmö and focuses specifically on cropping ecologies and they are apparently looking for researchers to come and experiment with it. I’m excited to connect with them and see what we might be able to do together!