When IoT Meets Agriculture: The Story of AGRIOT LAB Technology-Transfer Initiative

Raffaele Giaffreda
May 15, 2018

 

It is no question that meeting Sustainable Development Goals will be more and more the focus of the coming years. Within these expected targets, precision agriculture is set to play an important role: climate change will be putting more strains on our hydrogeological system; this in conjunction with increasing demands for food will require particular attention dedicated to the efficient use of agricultural resources, especially irrigation water.

Trentino is a mountainous region in the North of Italy where water scarcity has never been a big issue in the past. In 2017 there also was a combination of weather conditions including also a cold spell and a series of frosty nights that in late April hit the region causing considerable damage to crop yields[1]. 2017 was the driest on record since 1800[2] and lack of water was not only a problem for irrigation but also for activating some of the countermeasures for fighting the frost via tree sprinkler systems[3].

This is just a specific example out of many similar stories that have reported in recent years, showing us undeniable trends and urging us to become more and better equipped to deal with the effects of climate change in the agricultural sector. Not only draughts and water shortages but also earlier fruit trees blossoming and extreme weather conditions are poised to influence traditional crop management procedures. Issues of cost, in a traditionally low-margin business context, remain, as well as infiltrating technology in well-established tradition-rooted practices.

This is a domain where IoT can certainly help. Sensors and connected objects have made that transition from being of interest for high-margin applications (i.e. quantified-self gadgetry) to becoming so affordable and appealing also for low-margin markets such as agriculture. Furthermore, thanks to the relatively recent advent of low-cost LPWAN technologies, sensor data collection and wide fields coverage has never been easier and cheaper. Better quality, smaller size, more precision, reduced power consumption, these are all features that have considerably improved the IoT ability to create low-cost wide-coverage solutions. Sensing and M2M communication networks are also complemented by hundreds of IoT platforms to choose from.

Having said that, stitching together a reliable and useful end-to-end system in this peculiar application domain will require certainly more than just system integration knowledge and an IoT platform license, as it is emerging from AGRIOT LAB[4], a technology transfer activity where a well-established expertise in the IoT domain is used to address the requirements of an agricultural sector characterised by long-standing traditions and relatively high resistance against technology.

To smoothen the adoption path, it is essential to interact as early as possible with agriculture and forestry domain experts and ensure the problem being addressed with an AgriTech solution is matching what these experts would consider a real problem. In particular, part of the solution design was influenced by discussions with Laimburg Research Centre[5] and Beratungsring[6] experts in the apple orchard domain. Results in this field[7] published few years ago, show considerable irrigation water savings that could be targeted with the right sensing devices and knowledge: from 50% to 90% irrigation water use reduction in the region along the Adige river, notoriously fertile for this type of crops, without affecting the yield.

From interactions with other local stakeholders it also emerged that, reducing water consumption to fulfil the plant basic needs (rather than trusting the farmer’s perception) also brings reduced electricity bills as a side effect: in many areas on the hills along the main valleys the water must still be pumped up from lower altitude natural or artificial basins/reservoirs to ensure proper irrigation during dry spells. Through these interactions, the design of cheap, yet reliable sensing solutions was targeted, paying also particular attention to how easy it would be to deploy and use sensors, what would be the required maintenance and how much would be the overall cost.

The development of a worthwhile product and associated solutions was motivated by these market needs as well as by the willingness to ensure competitiveness of the local agricultural sector and a concrete impact towards SDGs. To achieve these goals also requires working together with agronomists to target the collection of the right data at the right frequency and from the most appropriate locations. All this know-how has crop-based variations and soil/climate dependencies which, once mastered can add to the expertise and to the ability to tackle this domain in many of its facets. This is a clear differentiator for AGRIOT LAB in terms of competitive advantage and in terms of the ability to scale up and create a sustainable business in this low-margin AgriTech domain.

Customer trust must be earned gradually and patiently, hence the need to start with something simple. Working together with customers throughout experimentation has revealed that an overwhelming technology-push would produce undesired resistance and friction against adoption. The first experimentation, therefore, targeted the implementation of a sensor for ground humidity measurement purposes (Fig. 1) and a wet bulb temperature sensor (Fig. 2) to generate alerts in case of frosty conditions. These are all easy-to-deploy, stand-alone wireless devices, with low-maintenance being addressed with a board design that paid particular attention to energy efficiency in data collection, processing, and transmission. The pictures show the humidity sensor, as easy to deploy as planting a stick in the ground and the wet-bulb temperature sensor, tied to one of the supporting poles present everywhere in commercial apple orchards. To meet the low-cost requirement, the large coverage was achieved using LoRa as an M2M communication technology (Fig. 3 shows details of the weatherproof gateway implemented for the experimentation).

Reliable hardware deployment must be supported by adequate software for secure data acquisition, interpretation, and storage. It is important to ensure one can provide flexibility in where the needed infrastructure is deployed according to requirements. While public clouds are appealing from a low-maintenance perspective, many SMEs in the agricultural sector still require a private solution entirely running within their premises and under their direct control. Security and data privacy is often perceived as a must in ensuring company secrets, historical data collection and field operations are duly protected in such a highly competitive market. One must, therefore, provide means to encrypt data during communication/collection phase at the hardware level (ensured through LoRaWAN), but also achieve security at the software level with appropriate device registration and control procedures and well-designed access policies for stored data. To provide this level of flexibility AGRIOT LAB relies on its own IoT Platform developed and validated in many application domains, but also on appropriate partnerships that have been developed throughout years of collaborative projects activities.

The ongoing experimentation phase is expected now to validate the solution and ensure the transition from lab-based prototypes to outdoor commercial solutions is properly tested before considering further enhancements and engaging in more sustained go-to-market activities. The direct involvement of end-users in the experimentation has been key to establish a trust relationship where constructive feedback is gathered and the foundations for future enhancements are duly laid. Infrastructure and field test activities will be completed with the participated design of a suitable graphical user interface for visualisation of alerts and actuation control. The subsequent GUI development will be the result of personalised interactions with different customers, to suit different needs tied to different cultures and crop specificities and further tailored to the sensors deployed and related data interpretation.

Figure 1: Tensiometer for soil humidity
Figure 1: Tensiometer for soil humidity
Figure 2: Wet bulb temperature sensor
Figure 2: Wet bulb temperature sensor
Figure 3: LoRa weather-proof gateway
Figure 3: LoRa weather-proof gateway

 

[1] https://www.munichre.com/topics-online/en/2018/01/spring-frost

[2] http://www.isac.cnr.it/climstor/climate_news.html

[3] https://fruitgrowersnews.com/article/protecting-your-fruit-from-frost-and-freeze/

[4] http://www.agriotlab.com

[5] http://www.laimburg.it/

[6] http://www.beratungsring.org/

[7] Thalheimer M., Paoli N. (2012). Bedarfsgerechte Bewässerung durch Einsatz von Sensoren. Besseres Obst 57 (6), 4-6


 

raffaele giaffreda2Raffaele Giaffreda received his first degree (Laurea) in Electronic Engineering from Politecnico di Torino in 1995 and his Master of Science in Telecom Engineering from University College of London in 2001. He joined British Telecom (BT) in 1998 as a researcher, shifting his interest from optical communications systems to data networks, mainly working on information retrieval systems and context-aware wireless networks. He has worked as Research Area Head for CREATE-NET since 2008 contributing to and leading many EU collaborative projects. He is currently Chief IoT Scientist at the OpenIoT Research Unit of FBK CREATE-NET working on IoT system integration, cognitive technologies and federated learning for IoT, use of blockchains for edge computing. He is also engaged in technology transfer activities in the AgriTech domain.