5G : a technological competitiveness lever for agriculture, but when and at what price?
written by Gwenaël Le Lay (CEO-FOUNDER / COPEEKS)
Since the announcement in 2020 of the auction of the frequency bands allocated to the so-called “5th Generation” mobile networks, there is not a week when someone explains to us, according to their “orientations”, that 5G will either be a revolution, a vector of productivity, an essential technology for the future or, on the contrary, a technical, ecological, economic or health nonsense.
However, far from all this media and political hullabaloo and the reeling effects of all kinds of experts, at COPEEKS, we deploy our solutions every day on farms in France and abroad and we are gaining experience and knowledge in the field concerning the use of mobile radio technologies.
The essential element in our eyes is to be able to use them to meet an identified use, a clear need expressed by our customers, for an operating cost that can be integrated into the economic model.
Also, even if 5G may appear interesting as a “concept/standard”, it is important to keep in mind that there is still a long way to go (5 to 10 years?) before it can be deployed operationally and effectively in digital solutions to support farms throughout the territory.
A question of coverage
The first questions we ask our customers when we deploy PEEK equipment with radio communication technologies (Wifi, 3G, 4G, LoRA): Do you receive your phone calls when you are in your rearing-buildings? Which operator do you subscribe to?
Indeed, in order to deploy an efficient digital service, it is necessary to have at least a link to the Internet at a reasonable speed (> 2G). However, the development and deployment of mobile antennas by national operators are linked to a number of potential users and traffic flows (the market) in a defined geographical coverage area. It is also clear that the majority of farms are far from urban areas, so their coverage by mobile antennas remains uncertain. Even if efforts are made (In France, the “New Mobile Deal” pact obliges territorial actors and operators by law to cover the so-called “white” areas), a densification of antennas is unlikely in these territories even with the advent of 5G. Indeed, this standard claims to use its frequency spectrum both to connect networks of connected objects (currently carried by SixFox and LoRaWAN technologies) and to carry massive data flows via ultra-high-speed connections. To achieve this, it will be necessary to deploy additional “high points” or “micro-cells” and to amortize these deployments with very high value-added digital services in order to establish a balanced economic model for operators. Therefore, in the investment plans of territorial actors or cooperative organisations located in low urbanised areas, the creation and management of “5G high points” can become an element of differentiation and territorial attractiveness in the next 5 years.
Adjusting the formats and volumes transmitted
The other element of importance when a geographical area is in low coverage is the composition of the data streams that the farm must transmit to the internet as a minimum in order to have the digital services to which the operator has subscribed.
In areas of poor coverage, anything that is ‘heavy’ in terms of data volume (video streams from cameras, remote screen streams (Teamviewer, PcAnyWhere, etc.), permanent connections, transfer of large files) should be banned.
To compensate for this low level of connectivity, it is necessary to work and think differently by analysing the data as close as possible to the place where it is collected (network engineering), to plan to archive and secure it locally and to transmit only the information that has “real added value” for the end user.
An example: several one-minute video sequences can be captured by a COPEEKS box. Instead of transmitting the acquired one-minute video sequence to the Internet (average size of the video file / HD quality: 100 Mbytes), it will be analysed immediately locally in the PEEK equipment using an approach known as “Edge Computing”. To do this, one or more neural networks residing in the PEEK equipment will deconstruct the video sequence frame by frame and identify the number of animals present in each scene, their posture (standing/laying down), the level of activity (amplitude of movements), the areas occupied (feed trough, refreshment area, etc.), etc. The “source” sequence will remain in the box, if it offers any value, or will be deleted after the analysis. The data resulting from the analysis (the added value) by the Deep Learning algorithms will be transmitted to the Cloud server (Cloud Computing) and as they do not weigh in volume (a few 100s of bytes at most), the transmission can be carried out on links with moderate speeds.
Through this more technical approach, the model becomes more virtuous as the carbon footprint is reduced, the evasion of raw traffic to the internet is limited, which leads to lower volumes of raw data outside the farm and controlled bandwidth consumption.
The same applies to data flows from sensors, weighing systems, water meters or the size of photos taken and transmitted — (today the vast majority of neural networks work on image sizes close to VGA format (640x480) to be efficient in processing time), so it is unnecessary in 2021 to transmit content in HD format
How much data should be transmitted outside the farm?
Today, the business model of all telecom operators is based on a monthly subscription fee that allows them to carry massive data flows (without too many limits) via increasingly powerful multimedia mobile terminals that interact in real time with service platforms operating in clouds spread around the world.
This model is not the most suitable for digital services offered on farms. Indeed, is it relevant to increase data flows to the internet, since the main user of the data is located on his farm most of the time. The essential information, the indicators modelled on the basis of raw information that is captured by probes, cameras, are in fact to be exposed to other third parties (advisors, technicians, veterinarians, etc.) and are to be sent to the internet. On the other hand, all the raw material that was used to create them — (several million data from sensors, several hours of video sequences, several thousand images) — does not need to be transmitted to the other end of the territory. Its archiving can be used in the future to re-train models or to develop new indicators.
Indeed, today, the computing power that can be embedded in digital systems (Embedded computing), makes it possible to process, analyse and enhance data locally, and to run one or more neural networks on these same data. The approach also concerns local archiving and security within the farm, thus limiting any escape of irrelevant traffic to the Internet. The term “Dataculturist” could then find its full meaning.
One way of using 5G in this context will be to move intensive calculations to the antennas closest to the farms, thus enabling the operator’s infrastructure to be used as an enabler of services and not simply a managed transport network. The economic model needs to be worked out but it makes sense.
Re-thinking the structure of buildings and the Farm Area Network concept
The other point to consider when communication technologies are used on a farm concerns the distribution of digital equipment that must communicate with the Internet (e.g. ventilation in poultry houses, milking robots, intelligent cameras in pig farms, weather stations in the field, RTK guidance for the tractor) and the level of connectivity required to transmit or receive all this information.
Today, each supplier/equipment manufacturer transfers the burden and management of connectivity to the farmer via a wired (ADSL) or satellite (e.g. NordNet) link, or sometimes 4G, so that he can use all the functionalities, update his system and benefit from a reactive after-sales service.
On a farm, it is not uncommon to find that several technologies and access points (ADSL, 3G/4G, WiFi) have been set up to communicate farm data to SaaS environments located several thousand kilometres away. The information is retrieved via graphic dashboards or mobile applications on the farmer’s phone, which cannot be consulted live from the farm due to a lack of signal. In fact, the data collected on the farm travels halfway around the world to be displayed and consulted locally on the farmer’s phone.
For the coming decades, the challenges of renovating farm buildings are often focused on thermal, welfare and functional aspects. Too rarely is the transmission and reception of data from the rearing-buildings considered. However, this is where the lever for productivity and competitiveness lies, particularly with the forthcoming arrival of robotic systems in the pig and poultry sectors, which have so far been poorly robotised. The materials and technical equipment used (metal ‘perfalu’ ceilings, reinforced concrete, motors, parasitic waves, etc.) are obstacles to the penetration of waves into the rooms and between the rooms, which strongly limits the use of digital solutions.
Before even considering installing 5G technologies in an agricultural area or territory, the essential question to be asked is that of setting up a telecoms network/infrastructure within the farm (as all “tertiary” or “industrial” companies do) by mixing the technologies best suited to the flows and uses envisaged (regulation control, monitoring cameras, calving detectors, alarms, faf, etc….), and make the installation completely secure by making it resilient in the event of a break in the main communication line.
If in the next 5 years, 5G covers this geographical area, then the operator will only need to switch its flows to this new network without any additional costs for the operation and be able to benefit from the numerous functionalities envisaged.
Further information: A file entitled “5G and agriculture” has been produced as part of the opportunity studies of the AgroTIC Chair. This file is available for download on the website (www.agrotic.org).
Access to the document: https://www.agrotic.org/wp-content/uploads/2019/10/5G-et-Agriculture_Janvier21_ChaireAgroTIC.pdf
This dossier is the result of documentation (from a wide variety of sources) and listening to various players in the agricultural and AgTech sector (including COPEEKS). It offers a vision, to date, of what 5G promises and the opportunity it could offer for the agricultural sector. It presents an analysis based on different agricultural use cases.
