This morning I had a chance to present our project “Self-powered sensor systems for Precision Agriculture” at the NTNU Faculty of Natural Sciences. It was great to get another perspective from plant ecologists about our research into improving fresh fruit and vegetable cultivation.
It turns out the wants and needs when it comes to sensors are not so different between plant ecologists and commercial agronomists. From both camps, there seems to be a lot of excitement around the possibility of using large networks of cheap, deploy-and-forget sensors.
Right now, temperature and moisture sensors can cost around $1000 per gateway and $100 per sensor. For researchers and growers alike, it is often difficult to defend the costs of more than just a few sensors. ONiO is aiming to reduce sensor system costs by an order of magnitude.
Having many sensors opens up a lot of possibilities. When there is a high degree of environmental variability, researchers can increase their number of study sites or replicates in each site. So too growers can get a better overview of variability within or between fields.
As well as upfront costs, there were questions from the plant ecologists around reliability. Some noted that they had difficulties with connecting sensor systems to 4G or with their built-in solar panels. They also explained they had to mail systems back in for costly repairs. Any sensor that is going to work in a study site or agricultural field needs to be built to last, unless they are so cheap that they are easy to replace.
Another issue discussed was accuracy. Will it be any lower for the cheaper sensors? As part of our project, we plan to mount the cheap sensors on top of the more expensive ones to see how they compare. We may also double up on cheap sensors to check their consistency.
There was further discussion around to what extent analytics based on sensors could be proactive versus reactive. It’s of course possible to react to feedback from sensors by adjusting greenhouse temperature and irrigation. Growers may also be able to proactively intervene if microclimatic variability is paired with local weather forecasts to, for example, map differential risk for pest and disease outbreaks.
Toward the end of our conversation we touched on the wider place of sensors in today’s food systems. Many modern plant varieties are bred to optimize yields over climate resilience. Sensor systems to detect localized flooding or drought will be needed to support them. Precision agriculture will also be necessary to minimize inputs and disturbance to build and maintain healthy soil biomes, a cornerstone of regenerative agriculture.
On our end, we were grateful for all the feedback. We’ll be keeping it in mind as we start our first field tests this winter in the greenhouse at Mære landbruksskole. We look forward to reporting back soon!
Norwegian University of Science and Technology (NTNU), Norges forskningsråd, Pedro Guimaraes, Vemund Bakken, Jenalle L. Eck
