Digitalización de la toma de decisiones en el sector agrícola a través de un sistema de gestión de información basada en Internet de las Cosas

  1. Juan Antonio López Morales
Supervised by:
  1. Antonio Skarmeta Gómez Director

Defence university: Universidad de Murcia

Fecha de defensa: 13 October 2021

  1. Miguel Ángel Zamora Izquierdo Chair
  2. José Santa Lozano Secretary
  3. José Cos Terrer Committee member

Type: Thesis


Based on the increase in world population and on the reduction in the available food resources, the optimization of farming methods needs to be improved in order to make better use of technological developments. These advances facilitate the execution of agricultural practices by transforming farms into efficient production systems that guarantee environmental sustainability. As farming becomes more intensive, having access to accurate information tailored to specific locations and conditions is essential to improve farm efficiency. Therefore, the sector needs to be prepared to capture, store and process large data sets from different sources and develop applications capable of responding efficiently to the sector's needs. The solutions developed must convert information into knowledge and facilitate farmers' on-farm decision-making, thus evolving from intuitive farming to scientific and intelligent farming. There are several obstacles that the agricultural sector has to overcome for the digitalization process to become widespread in order to reduce the potential drawbacks, so this doctoral Thesis has two clear objectives. On the one hand, to make sense of the enormous amount of data generated by the sector and the lack of analysis that makes it impossible to interpret it to improve its farms' performance. Furthermore, on the other hand, integrating all the information into an open and interoperable platform based on the Internet of Things (IoT) can improve data-driven decision-making for the agricultural sector. According to these requirements, the following objectives have been set: O1.To recognize and aggregate those datasets related to the management of natural resources of agricultural holdings. Deploy standardized architectures capable of managing them. O2. To analyze, adapt and improve existing information models and, if necessary, introduce new ones for the agri-food sector that facilitate data exchange and interoperability. O3. Use of an open platform following a layered model to allow, on the one hand, flexibility in its design and, at the same time, the modularisation of its components to enable different types of deployments depending on the smart infrastructure scenarios we wish to address. O4. Intelligent data processing for the extraction of patterns and models of the system under monitoring. These models will serve to provide predictive values and will form part of the sector optimization process. O5.Implement decision support tools for crops, irrigation water management, and energy efficiency, and the implementation of services associated with integrating data in its different technical, institutional, legal, and social aspects. The primary catalyst for this Thesis is the actual data used in the agricultural sector. The ability to express in an information model the concepts and relationships of the data of a farm is essential for the responsible technicians to clearly state what actions are allowed and forbidden to avoid misuse of information management. This is essential for agricultural technicians and service providers or other agricultural companies as it allows them to know which resources they can obtain and reuse to avoid duplication of data and generate mechanisms to improve the interoperability of the agricultural sector. For these reasons, the use of information models is considered beneficial to enable integrating any service, device, or characteristic element within the agri-food sector. The IoT-based architecture enables us to deploy systems that allow the fusion of data and the integration of data analysis procedures to improve the digitization of agriculture by creating new services that improve decision-making. An IoT-based architecture allows the integration of devices from different vendors, implementing open standards and interfaces and information models to improve decision making in the agricultural sector mapped to the NGSI-LD-based data model. In addition, it exploits the stored information to analyze factors associated with the production process, crop evolution, and optimal use of irrigation water on farms. The move to adopt NGSI-LD actively contributes to the growth of models for agriculture and represents an important step towards realizing the opportunity to generate a global market for solutions that provide intelligence to the agri-food sector. In doing so, the techniques applicable to the agricultural sector can be significantly improved and enable higher sustainable social, economic, and environmental returns. In addition, the range of services that the architecture can offer can be extended as a result of the level of interoperability achieved. Finally, it has been found that by integrating disparate data through the proposed architecture, services have been developed to improve the performance of daily agricultural tasks. As one of the primary energy consumers on farms are irrigation wells, agricultural pumping is monitored. For this reason, energy indicators have been integrated to prevent consumption in high tariff periods and improve the maintenance of the installations, taking into account the changes that are going to take place in agricultural electricity billing; these advanced monitoring mechanisms will enable farms to reduce their costs. Another element to be taken into account is marked by the climatic variability in which we find ourselves, which forces us to calculate the current conditions that determine the behavior of crop varieties in geographically different areas, specifically stone fruit trees. The design of a system that determines these conditions improves decision-making on the degree of climatic adaptation of the varieties to each area; this leads to an increase in the productivity of the trees and anticipates scenarios for early warning on the convenience of locating varieties in areas according to their cold requirements and the risk of frost.