Integrated Service Architecture to Promote the Circular Economy in Agriculture 4.0

Data digitization in agriculture begins with identifying the variables to be controlled when planting on farms [84], such as temperature, moisture, and soil nutrients, as well as weather conditions such as humidity, air temperature, light, crop nutrients [76], and pest control [121].

IoT enables this through sensors installed at various points on the farm, exchanging information between electronic devices [113] and applied in agriculture to provide monitoring of weather, soil thermal conditions [81], and water [29,30], pest and disease detection [103], and product traceability [101].

Industry 4.0 technologies applied to agriculture enable real-time data collection, input management, soil control, disease prevention, increased production, optimizing resources, and generating income [114]. The use of drones with IoT sensors is already a possibility for data collection in scenarios where the installation of sensor networks is not possible [91].

Temperature affects crop growth and the breeding of bacteria, and through a system of collecting air and soil temperature and humidity, a mathematical model was developed that evaluates environmental conditions and compares them with established standards, generating alerts for decision making by farmers [126].

The need to reduce the use of natural resources such as water and fertilizers in agriculture, the treatment of waste generated in planting, harvesting, and food industry processing activities, and the increase in production [14] have been the driver for the deployment of Industry 4.0 technologies in agriculture [117].

The analysis of climate conditions for agriculture is based on three factors: resource management, ecosystem conservation [114], and adequate services to farmers using information technology, thus generating profitability, product quality, and reducing the environmental footprint, with less use of pesticides [90].

Crop irrigation uses a source of water that is transported through pipelines and with on-and-off actuating devices and valves that then control the flow of water. The amount of water depends on each crop, and several forms of irrigation can be used, such as drip, sprinkler, furrow, and manual [127].

A simulation of irrigation control applied on farms in Brazil, with the use of IoT technologies, is a model considering the parameters of water requirements and irrigation time, which directly affect water use efficiency and evapotranspiration [30]. On the other hand, a low-cost system with environmental data acquisition can provide measurements of air temperature and velocity, soil temperature, pH, and dissolved oxygen, and thus control plant growth and environmental impacts [26].

A potato production system in Cyprus [114] was developed by using a sensor network to collect air and soil characteristics, and the data were sent to the cloud, where they were stored and analyzed, thus providing services to farmers for pest control and commands for automatic irrigation and fertilization [95].

Climate change has a strong impact on agriculture, and parameters such as water scarcity, soil degradation, increased energy needs, population growth, and increased demand for food impact the search for solutions [114]. For instance, the construction of smart greenhouses makes it possible to manage environmental variables [113] with control of humidity, air quantity, temperature, and favorable characteristics for plants [28,111] using models to manage heating, ventilation, CO2 control, and artificial lighting in an infrastructure that contributes to precision, sustainable, and manageable agriculture [113].

The data collected by the various sensors dispersed throughout the crops need to be sent for storage and analysis to extract information that can assist in decision making and return automated commands for control, since the Internet is not yet fully available in agricultural areas. It becomes necessary to use wireless network technologies in agriculture digitization projects, collecting data from distributed sensors in the field [4,84].

Applications with wireless sensor networks in smart agriculture include covering large areas with interconnecting sensors [6] and gathering large amounts of data. Activities in agriculture have integrated IoT technologies in robotic systems and been applied in planting, harvesting, and food production to enable cooperation between workers and smart systems and information sharing in the food production chain [107].

A proposed innovative smart IoT-based system based on microprocessors and a Single-Board Computer (BSC) for data logging using various sensors [50] can monitor soil moisture, air humidity, air temperature, and UV light, supporting precision irrigation of crops and enabling real-time exploitation of data for minimization of errors, as well as forecasting.

In Colombia, on a coffee farm, wireless sensor networks are used to collect leaf condition data remotely through image analysis processing, detect diseases on the plantation [94], and generate information for pesticide application. This strategy decreases the risks of crop loss and improves the productive performance of the farm [112].

Automated fertigation [109] is applied in precision agriculture by collecting data on moisture and soil characteristics, and through an IoT system and sensor networks with ZigBee technology, the system was developed to control fertilizer use and reduce water consumption [14], thus providing reduced environmental impacts [95].

The growth of the number of sensors scattered in agriculture can generate a large amount of data, and latency problems may occur in sending data through technologies such as Bluetooth in relevant virtual network scenarios [75]. Applications using machine learning [93] with a prototype for acquiring data from multiple sensors, machine learning algorithms, Geo statistics, and localization of IoT devices [115] are examples. The case study using computer vision [94] and machine learning [27] for inspecting coffee leaves on a farm in Colombia is an important practical example.

The use of tractors in agriculture [3] plays an important role in farms and [98] modern combustion systems with electronic systems, including the use of algorithms and artificial intelligence [78], managing fuel injection, reducing pollutant emissions, lowering fuel costs [5], and improving environmental and economic gains.

Tractors equipped with GPS (Global Positioning System), through the CAN-BUS (Controller Area Network) protocol, transmit the data of tasks performed and, after analysis, evaluate the most appropriate use of the tractor, extending the useful life and reducing maintenance costs [87].

In research conducted in Brazil, agricultural machines equipped with automation technologies have proven very productive in sowing, spraying, fertilizing, and harvesting activities, doing so autonomously, with relevant economic gains and lower cost per hectare [3].

An automatic irrigation control system using IoT, with data input from weather stations and soil moisture sensors interconnected by gateway LoRa (Long Range) networks and command messages sent by the MQTT protocol (Message Queuing Telemetry Transport) [6], adapts the efficient use of water on a plantation, generating savings and better product quality [8].

It is important to note that using advanced technologies such as IoT and AI to manage irrigation is a way to maximize crop yield while minimizing water consumption, in line with Agriculture 4.0 principles [51].

Nonetheless, intelligent sensing systems based on the edge-computing paradigm are essential for the implementation of Internet of Things (IoT) and Agriculture 4.0 applications, and the development of edge-computing wireless sensing systems is required to improve sensor accuracy in soil and data interpretation [47].

The Farm Scenery layer is presented in Figure 3, and the environmental variables to be collected are environmental humidity, temperature, luminosity, temperature, pH, water flow, and soil parameters, as shown in Table 2.

After the stage of data analysis using computational tools and the extraction of information, the next stage is the output, with the visualization of information through graphical panels available on web pages or mobile devices such as smartphones, printed reports, and commands of action in devices installed on the farm, making some processes automated. The use of software for agricultural planning provides management of farms, the production chain, and the distribution of the food produced [121].

The monitoring of environmental variables is context-sensitive and provides information to end-users such as farmers, agronomists, and traders for decision making and supporting agricultural production [16].

Machine learning plays an important role in data processing, collection, extraction, and mining activities using algorithms, neural networks, and artificial intelligence, which are applied in a variety of areas, such as energy, transportation, mining, shipping, healthcare, banking, security, and agriculture. Security requirements using machine learning in industrial systems [97], which analyze Blockchain-based strategies to preserve privacy in industrial systems with data security in sensors [19], can lead to product traceability and automated logistics management, saving time and financial resources [108] and decreasing food waste [88].

The application of machine learning and analysis techniques in agriculture and livestock provides productivity gains with the management of animal health and flock management, thereby improving land use, controlling animal nutrition, and reducing the environmental impacts generated [96].

The use of technologies in data processing in agriculture offers opportunities for the integration of diverse systems. A review of the technology standards used in cyber-physical systems (CPSs) [125] applied in advanced manufacturing builds the ontology of the 5C architecture layers (connection, conversion, computation, cognition, and configuration), thus enabling the integration of sensors, actuators, and protocols and their application in various areas. In agriculture, CPSs are used in planting and harvesting equipment, collecting information on soil conditions and water resources, and storage in cloud systems for further processing [107].

The use of IoT platforms with microservices involves independent processes acting on specific activities, such as, for example, data collection of soil, climate, and parameters of irrigated rice and cotton crops [110]. The platform involves hardware and software in the cloud, and focuses on irrigation management, presented scalability, flexibility, robustness, security, and performance, with possibilities of commercial applications in other agricultural crops.

An infrastructure using low-cost IoT and free software code to monitor weather conditions in organic crops, in a greenhouse, automatically controls and generates information made available through graphs on web pages with HTML protocol (HyperText Markup Language) [28].

The organization of farms by management zones [109] reduces costs for fertilization, pesticides [82,83], and water, and with the agricultural records of each one, one can have a decision more suitable to the needs of the crop. This strategy applied in a vineyard in Spain allows the crops to be classified into three levels of interest: soil, plant, or product, thus creating a map with overlapping layers to determine wine quality [25].

Sugarcane is planted in several fields on a farm, and as the plantation matures, fields are selected for harvesting, which is called harvest windows. A model to guide planting and harvesting on farms in South Africa [122] caused increased productivity and optimization based on historical data collected when adjusted in real-time. A machine routing system is possible in sugar production using sensors and IoT techniques, replacing labor and providig more suitable routing, higher repeatability, decreased fuel consumption, and better process control, with economic and environmental gains [116].

The management of agricultural supply chain activities, with the adoption of ERP (Enterprise Resource Planning) systems and BI (Business Intelligence) techniques, enables communication between those involved in the agricultural chain [105], more adequate control, faster solutions [7,104] to respond to climate variations, pest control, crop harvesting and management, and market trends [118]. Ref. [121] presents studies among farmers in Maranhão, Piauí, Tocantins, and western Bahia, in soybean, corn, cotton, coffee, sugarcane, beans, and fruit crops where production and marketing integration, material flow automation, data management, and more accurate diagnostics to support strategic decisions have been adopted as tools of smart farm managers.

The presentation of information in graphical dashboards becomes a decision support tool for farmers, with the use of Internet environments and mobile devices making access much easier and more intuitive [7,28].

The Information Processing, Analysis, and Output layer is shown in the Figure 5.

The substantial volume of data generated through the utilization of IoT in agriculture is stored via networks and cloud computing. Leveraging Big Data and analytics strategies, it facilitates enhanced production by aiding decisions concerning planting, harvesting, and production, and automatically issuing commands for irrigation, thereby saving the scarce resource of water [5].

A system facilitating the collection of crop data, including air temperature, humidity, and soil temperature, stored as input variables in the cloud, has undergone analysis to derive actionable insights for controlling corn crop cultivation [16].

Utilizing a wireless network architecture to connect sensors and a LoRa communication gateway for data transmission to the cloud, the subsequent step involves data analysis and extraction of information for decision making. This process also includes sending commands to regulate irrigation valves, fertilizing devices, lighting control, and heating/cooling systems, embodying an autonomous application [16].

A prototype for autonomous irrigation, developed and deployed in a farmer cooperative in Spain, gathers data on air temperature, humidity, and soil characteristics. These data are transmitted via a LoRa network with a coverage range of 5 km to a cloud server. Subsequent analysis returns commands for irrigation control, while mobile applications provide information to farmers [100].

In Indonesia, image processing using machine learning algorithms was applied on a cocoa farm. Texture data of the beans, collected remotely and transmitted to the cloud, underwent classification using artificial intelligence techniques, demonstrating superior performance compared to traditional visualization methods [106].

The processing of vast sensor-generated data on farms is routed to the cloud [114,119]. Leveraging Big Data tools [85,117], these data are analyzed and converted into actionable information for managerial decision making and issuing automation commands for field equipment.

A decision support system incorporating AI and a suite of machine learning algorithms aids in enhancing overall crop harvest quality and accuracy in precision agriculture [52]. This research utilized a dataset downloaded from Kaggle, containing eight features with seven independent variables, including N, P, K, temperature, humidity, pH, and rainfall. Furthermore, optimization techniques were employed to further enhance performance in smart factories.

Figure 6 illustrates the complete design of the Integrated Service Architecture derived from the results of the systematic literature review of selected articles, aimed at promoting the Circular Economy in Agriculture 4.0.

In the layer designated as “Information Processing, Analysis, and Output” within the Integrated Service Architecture aimed at advancing the Circular Economy in Agriculture 4.0, the references are categorized according to their respective applications, including irrigation control systems, fertilizer and pesticide application, crop management, greenhouses, and organization and management. These classifications are presented in Table 3 for clarity and ease of reference.

The integration of Industry 4.0 technologies in agriculture empowers farmers with control over crucial resources such as water, nutrients [109], pesticides, energy, machinery, robotic devices [77], and human resources [32]. Data-driven decision making and innovative business models [119] are instrumental in enhancing production strategies [13], reutilizing process waste [114], and minimizing losses across the agribusiness chain.

Social benefits encompass improvements in food security [1], reduced dependency on hazardous inputs, and the creation of employment opportunities in rural areas [32]. Initiatives fostering inclusivity and community engagement contribute to the formation of social capital and overall societal well-being.

Important resources for food production are wasted, with only 40% of irrigation water reaching the plants, and merely 5% of the applied fertilizer being transformed into nutrients absorbed by humans. Soil degradation affects between 30 and 85% of agricultural land, exacerbating food insecurity issues [39].

Following the linear production model for food production is no longer sustainable, especially in a scenario where necessary resources are increasingly scarce [128]. Therefore, to minimize the utilization of finite natural resources, it is imperative to adopt the principles of Circular Economy, which include preserving and enhancing natural capital by managing finite stocks and balancing renewable resource flows, optimizing resource yields through the circulation of products, components, and materials in technical and biological cycles, and fostering system effectiveness by addressing and mitigating negative externalities [39]. Figure 7 illustrates systems of the Circular Economy in a diagram, showcasing the management flow of renewable energy and finite materials, while highlighting the fundamental principles of circularity.

In the systematic literature review on Agriculture 4.0 and the associated technologies, only two articles were identified that mentioned the Circular Economy, along with a study outlining the barriers to its adoption [84], highlighting the lack of incentive and governmental support. Main challenges identified include pesticide usage and unproductive laborers [83]. Although the Circular Economy was not explicitly referenced, several articles demonstrated evidence of its application. Therefore, based on the ReSOLVE framework, evidence of Circular Economy implementation in Agriculture 4.0 was identified. This includes models for Regenerate (soil regeneration, nutrient recovery, utilization of renewable energy, and finite resource reuse), Share (equipment and technology sharing, as well as waste sharing), Optimize (automation, pesticide and water usage optimization, and energy consumption optimization), Loop (remanufacturing, reverse logistics, and recycling), Virtualize (information virtualization and remote service utilization), and Exchange (integration of new technologies such as IoT, cloud, Big Data, and robotics). The ReSOLVE framework operationalizes these principles through six key actions: Regenerate, Share, Optimize, Loop, Virtualize, and Exchange, as shown in Figure 8.

These actions involved analyzing the articles to ascertain the evidence of Circular Economy usage, which was then compiled and is presented in Table 4.