Mendel, data at the service of nature
Incorporation of technology in the agri-food sector is becoming increasingly more important for the competitiveness of agricultural producers. The market demands products of higher quality and better presentation, meaning production systems are in constant renovation. With modern agricultural techniques, such as hydroponics and aeroponics, plants can grow without soil using a nutritive solution. However, maximising and optimising crops in a greenhouse requires data collection at different points. That is due to the distributed environment, where climatic conditions vary depending on location.
Nowadays, the use of wireless sensor networks is expanding exponentially while fuelling the fourth industrial revolution and revolutionising the agricultural sector. A new wave of inexpensive, low powered wireless sensors has created perfect technological conditions for the agriculture environment, where long distances need to be covered, usually without access to electrical supply. The evolution of the technological offering for agriculture is evident by comparing the current market landscape with just ten years ago. Back in those days, only a set of unitary and costly sensors of various types were used to monitor an entire greenhouse, obtaining one data type for each variable, such as temperature, humidity, solar radiation, and the data was assumed unitary for all of the infrastructure. Over the years, technology has become accessible, and new communication protocols have emerged, making data acquisition and collection more efficient and accessible.
There are several sensor connectivity alternatives in the market to develop a monitoring network in agricultural facilities. The scale required, and acquisition and maintenance costs are usually the main factors when considering a new implementation. The cellular network, or alternative operators focused on the connection of M2M sensors, are a popular option. Those alternatives require highly qualified staff to perform sensor integration, as well as the associated maintenance cost. It is possible, however, to use networks without a predefined infrastructure, which provides device mobility without impacting communication and an effortless installation, while keeping maintenance costs low.
Thingbook.io and Zerebro Systems implemented Mendel, a cost-effective solution for intelligent monitoring, predicting and recommending actions for hydroponics and greenhouses. With Mendel it is possible to multiply system nodes or discard them in the event of an operational anomaly. Mendel uses RF24 modules, which operate in the 2.4 GHz free band in Europe. The cost-effective connectivity and maintenance of Mendel allow a transparent escalation in the number of nodes with sensors.
Want to add a caption to this image? Click the Settings icon.
An increase in the number of nodes means an increase in the amount of data collected. With more data, more metrics are available to analyse, and more automated decisions can be taken to proactively improve operations. To handle increasing amounts of data, Mendel uses T-Turing. T-Turing is a cloud-based system developed by Thingbook.io, which implements Deep Learning (Convolutional Neural Networks) and Machine Learning. T-Turing enables the detection and labelling of patterns from physical and logical sensors in greenhouses and other agricultural facilities. It can discover anomalies in real time, providing Root Cause Analysis for anomalies with an "Anomaly Signature", as well as making predictions for any sensor or group of sensors. Environmental data, water quality and plant status, among others, are sent to T-Turing Cloud, where farmers and professional users of the sector system can see and use both predictions and recommendations, such as the pattern of needs and anomalies detected.
At the same time, T-Turing cloud is used by a coordinating node to control the different automatic actuators, such as fertiliser dispensers, lights, water valves. The graphical and control interface provides access to data analysis results and recommendations, as well as enabling remote interaction with sensors and actuators. The system has been deployed in different greenhouse farms across Mexico, with different greenhouse conditions, as well as in hydroponic farms.
For more information about the system, contact firstname.lastname@example.org