Decarbonizing agricultural machinery is essential for achieving climate neutral energy systems, yet evidence on tractor energy demand remains limited, particularly for specialized tasks. This study analyzes 717 h of Controller Area Network (CAN) data collected from 75 kW narrow diesel tractors operating in vineyards to assess feasibility and environmental performance of electric alternatives. Measured duty cycles were used to quantify mechanical workload and power patterns across 14 representative vineyard activities. The operational feasibility of battery electric tractors was evaluated for battery capacities ranging from 36 to 300 kWh. Greenhouse gas (GHG) emissions were also estimated for multiple electricity supply scenarios, including European Union (EU) electricity mix in 2020, EU grid decarbonization scenario for 2030, and photovoltaic or agrivoltaics systems. Results show that vineyard tractors operate at partial load, whereas a limited number of energy intensive operations drive peak power demand and battery requirements. Battery capacities of 36-60 kWh enable completion of only 17-39% of observed operations, highlighting the limitations of small battery configurations. Larger packs improve feasibility by up to 98% of operating hours but peak daily demand can exceed 500 kWh, indicating the need for charging or battery swapping. Regarding emissions, the use of electric tractors charged from grid energy would reduce GHG from 24.66 kg CO2e h-1 of diesel to 16.09 kg CO2e h-1 of electric tractors recharged using grid energy given by European 2020 electricity mix and to 10.898 kg CO2e h-1 for the EU 2030 expected energy mix, whereas coupling electrification with photovoltaics would lower them to 1.95 kg CO2e h-1. Future work could expand the assessment to full life-cycle impacts including batteries and charging infrastructure, while research on transmission architectures could focus on decoupling equipment power demands from motion power demands.

Agriculture is one of the most hazardous occupational sectors, with conventional safety training often suffering from limited realism, low engagement, and reduced opportunities to reproduce hazardous scenarios safely. This article provides a structured analysis of digital and technology-enhanced approaches for safety education and training in agriculture. A literature-based framework was developed by combining five groups of digital tools—immersive technologies, simulators and digital twins, serious games and gamification, e-learning systems, and AI-based educational tools—with five learning contexts: formal education, vocational education and training, apprenticeships and traineeships, lifelong and on-the-job learning, and peer or community-based learning. In addition, a heuristic Application Priority Score (APS) was proposed to assess each technology–context combination according to evidence availability, cost–benefit feasibility, technology–learner fit, and inclusivity/accessibility. The results show that evidence is unevenly distributed, with stronger attention to immersive technologies, simulation-based approaches, and game-based learning in formal education, whereas lifelong, on-the-job, and peer/community learning remain less explored. The scoring framework also indicates that no single digital tool is universally suitable: the most promising solutions depend on the training context, learner experience, implementation costs, and accessibility barriers. The study provides a context-sensitive framework to support future research and the design of more inclusive and effective digital safety training solutions in agriculture.

Climate changes and their negative effects on the environment and society are now well known, also due to the increasingly frequent news episodes reporting the occurrence of extreme events with disastrous consequences in economic terms and in terms of human lives. Fires have and will have a major impact on agricultural resources and urban settlements, with critical consequences for the safety and health of citizens, the safeguarding of economic assets and the provision of essential services from fire-damaged ecosystems. Nowadays, satellite monitoring platforms are largely diffused. Therefore, environmental data start to be broadly available. This very preliminary work aims at proposing an effective workflow for wildfire prevention using the leaf area index (LAI) as precursor. By looking at the time evolution of three indices (mean Green, NDV and LAI) on an area interested by wildfire, we show that LAI has maximum before the fire and can be used as an indicator in defining a fire-risk category map. Though many more data is necessary to validate the method and ground truth LAI is lacking, the workflow presented could be the first step of an effective towards a fire-prevention system aiming at reducing the reaction time of safeguards and fire-damage.

Tractor rollover caused by physical contact with static obstacles or harsh terrain is a major hazard in agriculture. In addition, obstacle detection systems based on cameras and lidars struggle to perform well in agricultural working environments while their success rates might also be affected by a high number of false alarms. On the other hand, modern tractors are increasingly getting equipped with embedded GNSS and real-time kinematics (RTK) technologies that allow the driver to know its position on Earth with little error: this capability, combined with precise speed estimation which is a common feature of such systems, can be used to improve the safety of agricultural vehicles if static obstacles are georeferenced and adaptive buffer zones are created to serve as danger areas. In this study, which is part of a project called “Obstacle detection and tracking system for fixed and moving obstacles in agriculture” (SIRTRAck) funded by the Italian National Institute for Insurance against Accidents at Work (INAIL), a Keyhole Markup Language (KML) file containing georeferenced information on static obstacles in an experimental field has been created and passed to an algorithm on board of a tractor equipped with RTK, which integrated the dataset with the vehicle's real-time position, direction and speed. As the tractor entered the field, risk zones have been calculated by the algorithm from real-time motion information and have been used to alert the driver if the tractor entered one of them. The possibility of relying on ground truth and on adaptive shapes for danger zones built on speed, driver’s reaction time and expected slippage proved to be a valid and low-cost method for preventing tractor rollover hazards and might be easily used to improve existing geofencing features in both manned and autonomous agricultural vehicles.

Establishing a safe work environment in agricultural sites is a critical challenge due to the dynamic and hazardous nature of the scenario and the presence of large machinery. Effective tracking of workers and other obstacles around the machinery is fundamental for preventing accidents and improving overall safety. This paper presents SIRTRACK, a tracking system that integrates Light Detection and Ranging (LiDAR) and Ultra Wide-Band (UWB) technologies to track workers and obstacles intended to be installed at the farm machinery side. This sensor fusion approach leverages the complementary strengths of both technologies and is herein validated in a laboratory context.

Roll Over Protection Structures (ROPS) are conceived to provide passive protection for operators of heavy, self-propelled machinery in the event of a rollover. The verification of ROPS requires the application of a sequence of loads and the subsequent analysis of any resulting permanent deformations, that must guarantee access by the operator to an adequate living space (clearance zone). In many instances, the construction of structural components necessitates the iterative development of multiple physical prototypes. The utilization of computer-aided design (CAD) and finite element analysis (FEA) enables the visualization of the product and the evaluation of the mechanical resistance of the structure, even prior to the manufacturing and testing. This results in a significant reduction in the costs associated with the certification of the structure and the creation of physical prototypes. The aim of this study is to develop, by means of virtual prototyping tools, a driver’s cab for a self-propelled hazelnut harvester that complies with the OECD Code 4 standard for the approval of protective structures. Design for manufacturing and assembly criteria have been used for the 3D CAD modeling of the sheet metal assembly. Non-linear structural analysis was set up to simulate crushing tests, considering the elastic-plastic behavior of the material. Results and visualization of stresses and displacements show that the protection structure passes all acceptance conditions according to regulations. The designed driver’s cab has become an optional feature that can be installed on these vehicles, at the customer’s choice in place of the ROPS.

Stereo cameras, also known as depth cameras or RGB-D cameras, are increasingly employed in a large variety of machinery for obstacle detection purposes and navigation planning. This also represents an opportunity in agricultural machinery for safety purposes to detect the presence of workers on foot and avoid collisions. However, their outdoor performance at medium and long range under operational light conditions remains weakly quantified: the authors then fit a field protocol and a model to characterize the pipeline of stereo cameras, taking the Intel RealSense D455 as benchmark, across various distances from 4 m to 16 m in realistic farm settings. Tests have been conducted using a 1 square meter planar target in outdoor environments, under diverse illumination conditions and with the panel being located at 0°, 10°, 20° and 35° from the center of the camera’s field of view (FoV). Built-in presets were also adjusted during tests, to generate a total of 128 samples. The authors then fit disparity surfaces to predict and correct systematic bias as a function of distance and radial FoV position, allowing them to compute mean depth and estimate a model of systematic error that takes depth bias as a function of distance, light conditions and FoV position. The results showed that the model can predict depth errors achieving a good degree of precision in every tested scenario (RMSE: 0.46–0.64 m, MAE: 0.40–0.51 m), enabling the possibility of replication and benchmarking on other sensors and field contexts while supporting safety-critical perception systems in agriculture.

Noise and vibration are important risk factors for agricultural operators. In fact, musculoskeletal disorders of occupational origin linked to exposure to vibrations in the workplace have continuously increased in recent years and are followed, in terms of reported cases, by pathologies of the auditory system linked to noise exposure. In agriculture, the use of machinery for various cultivation operations involves exposure to these risk factors at levels that are often higher than the action values defined by the laws transposing European directives. In almond cultivation, as in other crops, harvesting is one of the most mechanized operations. The objective of this research is to carry out a comparison on workers’ exposure to noise and vibrations during almond harvesting organized in three different harvest yards: 1. Harvesting with a FACMA Semek 1000 self-propelled harvester; 2. Harvesting with Vimar MK2 shaker operated by John Deere 5105 GF tractor; 3. Harvest with Berardinucci Yellow Devil Vini 260 P self-propelled harvester. The average equivalent noise levels detected in the field, during harvesting in farms in central Italy, were respectively 84.6, 79.2 and 83.2 dB(A) with very similar maximum peak levels (137.7, 137.3 and 137.9 dB(C) respectively). The weighted accelerations transmitted to the whole body, detected on the most stressed axis (mainly y in yard 1 and x in 3, while in 2 the two axes are equally stressed) were, respectively, 0.29, 0.40 and 0.24 ms−2. The results show how the factor that most penalizes workers’ exposure is noise and, in particular, the peak values that exceed the upper action value of 137 dB(C) and which entail the application of specific prevention obligations towards exposed workers. Even with regards to exposures for prolonged periods, it is the noise which, in the case of harvest yards 1 and 3, can cause the action values to be exceeded, in the case of use of the machines for daily times exceeding 435 and 305 min respectively.

Recent years have seen a growing interest in hazelnut cultivation, justified by a doubling of the area cultivated with this crop since the year 2000. As a result, mechanization in this field, particularly with regards to harvesting machines, has also evolved significantly. Nevertheless, spread of electronics and telemetry systems only now has taken place, and there is a need to production mapping system. The aim of this work is to prototype and test a low-cost system for measuring the volume of harvested fruits using distance sensors. In detail, ultrasonic sensors have been employed connected to an Arduino microcontroller. Measured data have been acquired by serial communication and processed later using Matlab software. Several processing algorithms were tested in order to obtain more precise information about the volume occupied by fruits from the measured distances. Several tests were conducted by discharging hazelnut samples inside a sheet metal container that simulates on a small scale, the harvesting trailer. The container was provided with a lid, featuring holes for ultrasonic sensors fixing, and has been placed on 4 load cells, to monitor the weight of the hazelnuts during the tests. Measurements with ultrasonic systems show that the spherical shape of hazelnuts does not allow for high accuracy. However, the system allows adequate tracking of the filling status and distribution of hazelnuts in the trailer, and with the elimination of outlier data, it is possible to reduce the volume measurement error.

Performance analyses of mechanised vineyard activities require a reliable source of information that allows proper sizing of the tractor fleet according to field requirements and an assessment of the operating costs generated by them. To achieve that, however, a large amount of data regarding the power required by machinery and their field capacity, together with fuel consumption per hour and per hectare, is needed. This research, based on CAN-bus raw data collected from narrow tractors of 75 kW through a farming management information system (FMIS) system and then processed through a Python package called Vineyardutils created by the authors resulted in a dataset that summarises 374 labelled mechanised operations over a total of 717 working hours. The summary of each operation specifies its type, duration, idle time, speed and engine parameters such as temperature, engine speed, and torque; in addition, the dataset also includes the terrain slope, fuel consumption and size of the working area. As a result, the lowest values of fuel consumption per hour and per area have been estimated for activities such as fertilisation (4.95 ± 1.20 l/h, 3.73 ± 1.48 l/ha), whereas the highest values belong to harvest operations (14.70 ± 3.39 l/h, 11.69 ± 4.68 l/ha); regarding field capacity, values range from 0.54 ha/h for leaf removal to 4.46 ha/h for multirow crop protection. Furthermore, correlations have been found regarding environmental temperature and requested power, with evidence for most of the activities. Future developments can include different tractor engine setups and additional field data.

An overview is presented of the progress since 2021 in the construction and scientific programme preparation of the Divertor Tokamak Test (DTT) facility. Licensing for building construction has been granted at the end of 2021. Licensing for Cat. A radiologic source has been also granted in 2022. The construction of the toroidal field magnet system is progressing. The prototype of the 170 GHz gyrotron has been produced and it is now under test on the FALCON facility. The design of the vacuum vessel, the poloidal field coils and the civil infrastructures has been completed. The shape of the first DTT divertor has been agreed with EUROfusion to test different plasma and exhaust scenarios: single null, double null, X-divertor and negative triangularity plasmas. A detailed research plan is being elaborated with the involvement of the EUROfusion laboratories.

Overturning is one of the main causes of fatal accidents in agriculture. Farming activities that are affected the most by this type of hazard often involve trailers, balers, and sprayers: a tractor and any kind of towed equipment. This is a well-known issue, described by a great number of accident reports. But despite efforts in accident prevention the occurrence is still high and most of the assessments focus on addressing responsibilities and reconstruct the accidents by cause-effect relations. Accidents, in reality, are never the same because actually the working environment, social constraints, deadlines and seasonality play a major role in accident dynamics and therefore should be tracked by any safety assessment in agriculture as safety conditions gradually degrade with time or even in a matter of a single activity. The analysis of safety conditions through new methodologies and the realization of possible work scenarios are the objectives of this work.

Precision agriculture optimizes farming practices by minimizing resource use through site-specific measurements and treatments, with georeferencing machinery playing a crucial role for precise navigation and safety. The SIRTRACK project, funded by INAIL, aims to enhance unmanned agricultural tractor safety by integrating ultra-wideband localizers, LIDARs, depth cameras, and real-time kinematics navigation systems. This multi-technology approach addresses visibility challenges in agricultural environments and improves obstacle detection by cross-verifying data from multiple sources. The paper discusses the development of a communication system critical for transmitting RTK data from a base station to agricultural vehicles in the field, ensuring accurate positioning and operational efficiency. By selecting an appropriate radio propagation model, the study aims to determine which cost-effective and reliable communication device can be used to ensure a precise vehicle navigation and a long-term project sustainability.

Data-driven digital agriculture relies on the ability to gather, process and understand information from field operations. In this context, raw data generated by farming equipment and machinery represents a fair share of the overall information that results from agricultural tasks and can be obtained even by older equipment after proper retrofit actions. This possibility shall, on the other hand, be handled and filtered to avoid that erroneous data might worsen the quality of a whole dataset. The aim of this research, performed on a narrow tractor in vineyard mechanised activities, was therefore to evaluate a method for splitting raw data gathered from the Controller Area Network (CAN) bus into different clusters corresponding to specific tasks that can be carried out on the field such as tillage, fertilisation, shredding, treatments, defoliation and harvest. As the data could be associated to a specific task, further data mining showed possible errors, biases, noise and particular pattern in the dataset. As a result, a refined dataset of vineyard operations has been produced, plus additional information regarding tractor paths and idle times has been provided. Further analysis could provide state of the art knowledge on narrow tractors operating costs for each type of field activity in real agricultural work contexts.

Obstacle detection in mechanised agricultural activities is a key feature in safety management systems which prevents accidents involving workers on foot. Presence of obstacles in blind spot areas utterly represent a major issue for tractor drivers. This paper illustrates the model of a safety management system that aims to handle obstacles by using depth cameras and LIDARs to detect the obstacles and estimate their distances, risk maps that provide information obtained by Real Time Kinematics (RTK) georeferencing of fixed obstacles such as trees or buildings, assisted by an additional system based on Ultra-Wide Band (UWB) wireless communication devices for real-time localization of workers on foot and notifications through Bluetooth Low Energy (BLE) devices. The model is part of a project called SIRTRAck and also provides an assessment of a tractor's blind spots obtained by both field tests and simulations. The preliminary part of research activities showed that prototyping a low-cost aftermarket system for obstacle detection is possible by adapting different data acquisition methods for safety purposes and by making use of the information accordingly. Further steps of the project include a prototyping phase of a demo kit based on the model.

A modern method for learning to drive a tractor and learn how to manage its risks is the development of a virtual driving simulator, which allows both to become familiar with the instrument of guide, and to manage the various tools to work the agricultural fields. This simulator allows knowing how to meet or avoid hazards and above all, the risk of tractor overturning. Various driving simulators have been created, with relative successes but also with significant limitations, for this reason the Tuscia Department of Agriculture and Forest Science approach has different characteristics from those created previously. The gamification that uses amusement for educational purposes, is currently the best method for teaching and learning any context. The creation of a Tractor Driving Simulator with a Gamification approach requires both careful observation of the characteristics of the games, especially those relating to farming simulation, and the creation of a reliable physical mathematical model of an agricultural tractor. For the creation of this model, this topic proposes a method for sampling the behavior of true tractor in action on a real agricultural field. The objective is to find a 1:1 relationship between the real tractor and the virtual one, meaning that driving the simulator is like driving the real tractor, but also vice versa. The IoT-based method proposed, allows the creation of a consistent and economical database that will be the basis of the model.

Noise is a major physical hazard in agricultural activities, and numerous research activities have managed to detect its effects, resulting in surveys and measurements which help to define exposure limits, prevention methods, and control strategies. This review aims to collect and analyse the data from research studies and to provide a comprehensive overview on the subject. Thus, a set of 81 papers, gathered from the Scopus and PubMed scientific databases, has been analysed to provide information regarding the evolution of noise exposure levels over time, to highlight findings on noise-induced hearing loss (NIHL), and to list strategies for noise prevention and control in agriculture. Bibliographic research showed that noise measurements between 1991 and 2022, included in scientific research on farming, forestry, and animal husbandry, mainly reported values beyond the threshold of 85 dB(A); furthermore, several research activities on NIHL showed that farmers’ family members and children are often exposed to high levels of noise. Lastly, an analysis of the prevention and control strategies over time is provided, focusing on prevention programmes, screening, and the use of hearing protection devices (HPD). The identified literature suggests that additional efforts are required in regards to machinery design relating to the socio-technical aspects of agricultural activities and that side-effects of NIHL, as well as the negative impact of noise on other risks, might deserve further investigation.

The main goal of this paper is to set up the an adequate research program for the development of a low-profile, full-electric tractor model for orchard and vineyard. The prototype, properly arranged to support the agriculture 4.0 technology, will be equipped with a non-foldable and non-tiltable protective structure in case of overturning (ROPS), suitable to reduce the overall height of the machine, while effectively protect the driver in all working situations. To assure also a proper ergonomic level, reference was made to the European regulation 1322/2014; more in detail concerning safety, ISO 4254–1 and OECD Codes 4 and 7 were taken into account. In addition, to respect a front and/or rear track width less than 1150 mm (to make possible the classification as a narrow tractor), the prototype should have an overall height not exceeding 1600 mm and a minimum ground clearance of 250 mm. The research was initially aimed to the ascertainment of the tractor power required for the most common operations carried out in vineyards and orchards. As a reference, the CAN-bus data coming out of some high powered models (for the category, i.e. 74 kW) were analyzed, to investigate the real power requirement involving vineyard operations. The heaviest conditions were recorded in plant protection products distribution, carried out with the pneumatic sprayer. Moreover, a parametric analysis was conducted on a one degree-of-freedom 4WD tractor model to determine power needs for traction in different working conditions, concerning slope, travelling speed, terrain features, etc.

The work aims to analyse from a comparative point of view the legislation on health and safety in the workplace of the four European Member States with the highest number of accidents at work. The Member States chosen for the study are Germany, Spain, France and Italy which alone in 2021 recorded over 2 million accidents in the workplace and over 100 thousand in the agroforestry sector alone. First, the trend of accidents in the agroforestry sector from 2017 to 2021 was analysed: this analysis was fundamental to understand the trend of accidents in the Member States and understand the causes of the high numbers of accidents. Through a legislative comparison, we attempted to understand the similarities and differences in the various regulations of the sector in the Member Countries examined, trying to understand gaps and ideas for future improvement. The various legal structures, dual or single, of the four Member States were also examined, trying to understand which the most efficient and effective way is to limit accidents and deaths at work as much as possible, particularly in the agroforestry sector.