Back to the Future: What Is Trending on Precision Agriculture?

1. Introduction

The Industrial Revolution changed the way of cropping with new machinery. Tractors and agricultural implements created a new agricultural era. For over a century, the evolution continued steadily until the 1940s with the emergence of synthetic chemicals. Closer to the present time, the development of GMO crops increased global production but also the dependence on synthetic chemicals. Shortly after, the concept of Precision Agriculture was introduced. Precision Agriculture takes into account the spatio-temporal variability of crops. Thus, Precision Agriculture is a management strategy that gathers, processes and analyzes temporal, spatial and individual data and combines them with other information to support management decisions according to the estimated variability for improved resource use efficiency, productivity, quality, profitability and sustainability of agricultural production [1]. The availability of Precision Agriculture technologies has substantially increased. Sensors and actuators in agricultural machinery have increased crop yields while reducing inputs. Additionally, a new type of agriculture has emerged in the last decade. Precision Agriculture has also provided support for farmers in the supervision and implementation of agricultural practices in order to overcome the challenges posed by the perspective of feeding a growing population. Furthermore, Precision Agriculture has been reinforced by extraordinary technological advances that merged the physical, digital, and biological worlds. Technological innovations and the availability of new imaging tools and increased processing capacity of computing devices are significantly improving detection and classification processes. All these systems enable the variable rate by using the generated information. Thus, digital agriculture combines both physical and digital concepts in agriculture.

Current agriculture is experiencing a disruptive change, a new “industrial revolution” in which data management systems, variable application tools, satellite and UAV systems, FMIS, localized irrigation, robotic implements, artificial intelligence, and autonomous mobile robotics coexist [2]. The novel tools allow the early detection of pests and diseases and the hyperlocalized application of treatments while generating spatial information for a further decision-making processes. New robotic systems are replacing traditional machinery, and a new productive paradigm is emerging. The achievement of an ecosystem with human–robot interaction is becoming increasingly closer. The future of agriculture is now, and it is revolutionary.

2. Variable Rate Applications

Site-specific application allows a more precise treatment in agricultural fields, taking into account the spatial and temporal variations of crops, soil or pest populations [3]. This approach offers both economic and environmental benefits. When it comes to weed treatments, it can achieve this precision by applying treatments exclusively to weed patches (e.g., turning nozzles on/off) or by varying the application based on factors such as weed species composition, distribution or weed density [4]. A typical application system involves mapping the field and identifying the specific needs of the crops. With this information, the appropriate variable treatment can be applied by site-specific application machinery. For instance, if there are no weeds present, a less intensive mechanical weed control method or no application of herbicide can be performed by closing in that particular section or nozzle. Similarly, varying doses of fertilizer can be applied as required. Effective decision-making relies on comprehensive knowledge of the farmland and its heterogeneity. This knowledge is crucial for optimizing the use of agricultural resources and achieving sustainable yield potential. The agricultural industry is embracing digital transformation supported by various technologies developed in recent decades such as remote sensing, on-ground sensing, site-specific treatments, and decision support systems (DSS). These systems aim to enable variable site-specific application of treatments using modern technologies like monitors, ISOBUS and new tractor implements.

3. Digital Agriculture: Farm Management Information System

The development of a Farm Management Information System (FMIS) allows the capture and advanced analysis of the data generated by the agricultural systems installed in each machinery and vehicle sensor. FMIS represents the basis for any innovation in ICT at farm level, and integrates all the information of the current machinery as well as all the actions to be carried out in the immediate future by new developments. FMIS reduce production costs while maintaining high product quality and safety [5]. The collection, storage, analysis and exchange of agricultural data, as well as decision support, are the key to farm improvement worldwide. The interconnection of systems opens the door to improvement based on global experiences. Conventional FMIS currently consist of multiple standards, languages and protocols for different devices, sensors and information flows. This implies a lack of interoperability. Thus, new modular developments are integrating both predictive and descriptive models based on the Artificial Intelligence branch. The use of machine learning for the analysis of data from different sources in the processes generated in the field will allow automatic analysis and decision-making. This will allow FMIS to rely on Blockchain technology for true digital traceability.

4. Smart Water Management

The selective, effective and adapted dosage to the needs of each area seeks to maximize the performance of water application (cost reduction and improvement of recreational areas, parks, forest areas, etc.) in a way that positively affects the quality of the treated area and reduces the impact caused by each of the tasks and the work carried out [6]. This task needs different steps for its execution; the first of them is based on the knowledge of the variability present in the monitored area. Consequently, areas receive these specific treatments. Thus, management and analysis tools have to be developed to allow zoning these areas in an efficient way. The first step consists of detecting the needs and the creation of georeferenced maps that can be used by the equipment that will carry out the variable application. Mobile vehicles such as tractors, treatment machines or static networks with the capacity for zoning or localized action, as in the case of irrigation systems or specific fertigation, will provide a differential dose of both water and fertilizer. A large number of technologies have been developed for the differential application by means of equipment connected to machinery. On the other hand, although there are different commercial solutions that can be interconnected, there are functional software and hardware systems with protocols for analyzing the variability present and generating decisions for selective dosing. However, this solution is currently less developed.

Water management needs the analysis of data used for effective decision-making; it is of vital importance, and during the recent years, the availability of information has increased considerably. Satellite-image-based solutions that provide continuous information can be considered, as well as many other sensors that can be on board of the machinery or stationary networks that measure the hydraulic conductivity or other soil and crop values by different sensors. The information is sometimes complex to process, and it is necessary to generate simple and understandable tools to manage the information. Additionally, the Internet of Things allows us obtaining information in real time through sensors and different control elements. The interconnection of systems makes it possible to improve prescriptions and automatic decision-making on water management. Machine learning systems on hyperspectral image analysis, weather forecasting systems, soil sensors, plant sensors, etc., together with historical data and 5G technology are transforming agriculture and localized irrigation systems.

5. Agricultural Robots

The presence of robots is nowadays common in society. The industry has robotic solutions in all its production processes; even in our homes we rely on multiple robotic solutions that make our lives easier. These changes are also demanded in the agriculture sector. UAV (Unmanned Aerial Vehicle) and UGV (Unmanned Ground Vehicle) are already close to farmers in their daily tasks. The efforts made by workers and the sometimes-extreme working conditions require a profound change in the agricultural model. The use of exoskeletons or the development of collaborative robots is a research trend in the field of agricultural robotics. However, other types of robots and intelligent implements are already deployed in the fields. The use of mobile robotics has been widely explored, from universities to large agricultural machinery manufacturing companies. A common goal is the development of unmanned and remotely controlled mobile platforms, from aerial to on-ground units. Usually, the aim is to develop fleets that can work autonomously while the platforms are monitored in real time. The development of software for fleet management, simulation, tracking, data storage and information processing is the basis for these mobile robot fleets [7]. In this respect, there are multiple approaches based on robot sizes. While the use of multi-robot fleets is the most common, albeit a more complex, approach, large robots based on traditional platforms such as tractors are also present nowadays. On the other hand, fleets based on small autonomous platforms provide a number of advantages such as high adaptability, low soil compaction, safety for workers and the environment, substitutability in case of malfunctioning and scalability to different crop types and farm sizes.

Although there are multiple approaches to the future of agricultural robotics, all robotic platforms pursue the same objective: reduction in human effort, reduction in the use of resources, acquisition of crop data, improvement of the cultivation process, etc. This is currently changing the agricultural production paradigm. In the future, autonomous machines, intelligent implements, decision support systems, management platforms, etc., will coexist with traditional agriculture. The development of intelligent systems must consider the agricultural transition model in which the gradual adaptation of robotic technologies is already happening.

6. Artificial Intelligence

Artificial Intelligence systems integrated with robotics are ushering in a new era in the field of agriculture. The emergence of automated recognition systems, localized treatment systems, and manipulators connected to autonomous navigation systems brings forth a fresh prospect for eco-friendly systems that are more productive and capable of producing high-quality food accessible to all. Presently, it is evident that technological advancements will significantly transform agricultural production systems, enabling more efficient, precise, and sustainable execution of farming tasks [8]. Concepts such as digitalization, Information and Communication Technologies (ICTs), the Internet of Things (IoT), big data analysis, machine learning, sensors, Unmanned Aerial Vehicles (UAVs), agricultural robots, and agriculture 4.0, accompanied by Smart-Farming principles aligned with the circular economy, are increasingly being employed in rural areas. This trend facilitates the gradual adoption of Precision Agriculture strategies, addressing the issue of rural depopulation by introducing technological advancements to the agricultural landscape. Within this framework, crop monitoring for early detection of losses caused by pests, diseases, and weeds, and subsequently applying targeted phytosanitary treatments that cater to the crop’s variability are the two key applications driving the utilization of these technologies.

During the recent years, the use of machine learning has been evolving rapidly since deep learning is a form of machine learning that can significantly augment our understanding of intricate patterns by automatically expressing them based on simpler ones. Artificial neural networks like CNNs exemplify this approach. In this period, deep learning technology has achieved remarkable progress in artificial vision research, surpassing traditional artificial vision systems. For instance, it can learn to extract optimal features from input images, thereby reducing training errors.

7. Conclusions

Agriculture is undergoing a profound change nowadays. The emergence of new cultivation techniques based on data analysis, together with intelligent implements, autonomous platforms, information processing systems or integrated solutions based on fleets of robots equipped with perception, decision-making and action, are rapidly changing traditional processes and the way we farm. What science fiction imagined almost a century ago is now real. The development of an end-to-end robotic ecosystem is close at hand. There are many solutions on the market, and thousands of researchers and companies are investing their efforts into improving future solutions for smart agriculture. Robots will farm better to feed a growing population in which the use of resources has to be decreased and in which conservation of natural resources and soil conservation are a must. The large number of variables makes intelligent management agriculture together with mobility and crop treatment systems a clear necessity for farm management. However, the lack of standardization, interconnectivity of systems, and availability of information are still a barrier to overcome. The construction of a new agricultural model has begun, which in the coming decades will lead to a profound transformation of the agricultural management system as we know it today.