Conference Agenda

Addressing the escalating labor shortage in the fruit industry, the development of intelligent robotic harvesters capable of simultaneous internal quality assessment has become essential. Such systems enable the non-destructive evaluation of internal attributes without manual intervention. Near-Infrared (NIR) spectroscopy is a robust method for non-destructive Brix measurement; however, its spectral integrity in open-field orchards is often compromised by intense ambient sunlight and light leakage caused by the fruit's surface curvature.

To overcome these challenges, this study developed a novel robotic end-effector designed to create a localized darkroom environment on the apple surface. The device features a custom-designed structure to mechanically isolate the measurement site from external light while maintaining stable contact with the fruit. An integrated optical fiber connects the end-effector to a compact NIR spectrometer, allowing for the acquisition of high-fidelity spectral data by effectively mitigating solar interference.

Performance evaluations confirmed that the developed end-effector successfully masks ambient light and provides stable spectral measurements from the apple surface. This study demonstrates the feasibility of real-time quality grading during automated harvesting, confirming the technical viability of the integrated spectroscopic approach. Subsequent research will involve the integrating of the developed system into a robotic manipulator for comprehensive field validation.

Vertical farms enable precise environment control, but scalable automation requires integrated monitoring of crop growth and indoor conditions for robotic operation. This study presents an integrated control and monitoring system for autonomous agricultural robots in multi-tier vertical farms, combining real-time environmental sensing with AI-based image analysis. The system was implemented in (i) a rail-equipped strawberry farm, where a robot captures arm-mounted images while temperature, humidity, and CO₂ are logged every 10 minutes; and (ii) a multi-tier lettuce farm, where overhead cameras estimate volume and fresh weight using projected area and depth cues. YOLOv8-based models were developed to infer growth stage and harvest timing. The strawberry maturity detection model achieved 0.854 mAP₅₀ across five classes. For lettuce, the detection model achieved 0.995 mAP₅₀, and harvest-readiness prediction reached 95.24% accuracy. A unified dashboard visualizes ripe-fruit counts per line and predicted fresh weight together with time-synchronized environmental data. The proposed system supports environment-control optimization and harvest decisions, enabling efficient task scheduling and path planning toward fully automated precision agriculture in vertical farms.

The reduction of plant protection product (PPPs) use and related environmental impacts is a key objective of current European agricultural policies. Precision spraying technologies, particularly variable-rate and sensor-based systems, are a promising solution to improve application efficiency while limiting off-target losses. This study evaluates the performance of an autonomous robotic sprayer equipped with LiDAR sensors for real-time canopy detection and variable-rate application in comparison with a conventional air-blast sprayer in an apple orchard. Field trials were conducted in an apple orchard in Italy under three different canopy vigor conditions. Spray distribution was assessed using water-sensitive papers, while spray drift was quantified through dyed collectors positioned before and after the canopy. Results showed comparable canopy coverage between the two systems, with no statistically significant differences, indicating that reduced spray volumes can maintain treatment effectiveness. Although high variability prevented statistical significance, the robotic platform achieved an average drift reduction of approximately 50% compared to the conventional sprayer, with the greatest reduction observed in low-vigor canopy conditions. Moreover, the robotic system provided more uniform spray distribution. These findings demonstrate the potential of LiDAR-based, variable-rate autonomous spraying to reduce PPP losses and environmental contamination, supporting the development of more sustainable orchard management practices.

Nigeria, despite being a leading rice producer in West Africa, struggles to meet domestic demand due to persistent bird pest infestations. Traditional bird-scaring methods remain ineffective, causing economic losses and reliance on costly imports. To address this, researchers enhanced an existing mechanical bird scarer based on impulse and momentum transfer. The device, powered by a 2 kW motor, originally operated continuously at 9168 strokes/hr with a 30 m coverage area, but this led toenergy waste and reduced lifespan. The study introduced automation using an Arduino-based system with PIR motion sensors, which detect bird presence and trigger the motor only when needed. Programming was done in Microsoft C and C++. Preliminary field-like tests evaluated projectile velocity, launch angle, strike frequency, coverage, and landing dispersion using standardized stones at angles of 30°, 45°, 60°, and 75°. Results showed improved reliability, consistent strike intervals, reduced dispersion, and enhanced targeting efficiency compared to manual operation. Importantly, the system was built with locally available, low-cost materials, making it accessible to small-scale farmers. The study highlights the potential of affordable automation to strengthen crop protection, with future prospects including adaptive control and solar integration for sustainable rural deployment.

This preliminary study describes the development and initial technical evaluation of an intelligent system for the selective application of plant protection products in greenhouse vegetable crops. At this stage, the work focuses on canopy characterization and the generation of spray prescription maps based on crop structural parameters. The system integrates unsupervised artificial intelligence algorithms with LiDAR-based sensing to characterize crop structure through Plant Row Volume (PRV), used as a key indicator for spray dose adjustment. Georeferenced canopy data acquired by a LiDAR sensor are processed using dedicated software to generate prescription maps defining the appropriate spray volume. Two mapping strategies were defined: a uni-band approach, assigning a single PRV value per measurement section, and a multi-band approach, considering several vertical canopy bands per section to better capture canopy variability. Potential savings in plant protection products were estimated through simulation of a variable-rate spraying strategy. Preliminary results indicate substantial reduction, particularly under the multi-band approach, with estimated savings of 65-75%. Overall, the proposed technology demonstrates considerable potential to enhance application efficiency, reduce plant protection product consumption and support the objectives of the European Farm to Fork Strategy for sustainable agriculture.

Testing for quality machinery is an important component of sustainable agricultural mechanization. This paper provides a clear state-of-the-art overview and understanding of the requirements to foster testing activity in the Philippines. The Agricultural Machinery Testing and Evaluation Center (AMTEC) is the premiere testing center in the Philippines. As mandated by Law, all agricultural and fishery to be sold in the market shall pass through testing and evaluation of AMTEC. By the end of 2024, AMTEC had tested about 8,936 agricultural and fishery machinery. However, with limited funding, AMTEC could not hire the ideal number of personnel for the testing operations; could not sustain the upgrading and maintenance of testing equipment; and is not doing two (2) of its functions: field testing and publication of technical bulletins.

As ways forward, the following are recommended: a) testing program to assist local manufacturers; b) support research related to the testing and evaluation; c) development of an evaluation system in the selection of quality machinery; d) strengthened the establishment of strategic testing centers; e) support the Mutual Recognition Agreements for testing; f) tests to determine parameters other than machine performance; g) full inspection of the machinery; and h) the proper financing of AMTEC.

Mechanized harvesting is crucial for maintaining the physical and physiological quality of grains. Automatic harvester regulation systems have been incorporated into combines to adjust operating parameters in real time in response to varying field conditions. The objective of this study was to evaluate soybean grain quality as affected by an automatic harvester regulation system in mechanized harvesting, comparing operation with the system activated and deactivated. Grain samples were collected directly from the grain tank, separated into the following fractions: pure seeds, broken seeds, crushed seeds, whole pods, and plant and mineral impurities. Germination, viability, vigor, and mechanical damage tests were also conducted. A statistically significant reduction in whole pods was observed with the activated (0.10%) compared to the deactivated (0.32%). Plant impurities also showed lower values with the activated (0.19%) versus the deactivated (0.33%). No statistically significant differences were observed in physiological quality, with germination of 93.90% and 93.80%, viability of 85.90% and 89.30%, and vigor of 76.10% and 78.90%, for the activated and deactivated, respectively. The automatic harvester regulation system did not compromise grain physiological quality and contributed to reducing whole pods and plant impurities in the grain tank.

No-till systems have promoted significant advances in crop production by minimizing soil disturbance. However, the longitudinal distance between the cutting disc and furrow opener, combined with forward speed, critically influences the energy demand of seeding operations. This study determined the effect of cutting disc-to-furrow opener distance on draft force, drawbar power and fuel consumption of double disc and shank furrow openers at different forward speeds. The experiment was conducted in a 2×3×4 factorial design with furrow opener type, distance and speed as factors. Draft force was measured by a load cell and fuel consumption by a flow meter, while drawbar power was calculated from draft force and speed. The shank furrow opener demanded 57% higher mean draft force than the double disc (2.49kN vs 1.58kN). Drawbar power increased from 1.77kW at minimum speed and distance to 7.32kW at maximum conditions. Fuel consumption increased by 12% from the lowest to the highest speed. The double disc at 1.11m s⁻¹ and 0.50m distance represented the most energy-efficient configuration, while the shank at maximum speed and distance resulted in the greatest energy demand. Operational regulation proved more decisive than furrow opener type for energy efficiency in no-till seeding.

The emergence of hedge almond orchards is due to the faster harvesting with self-propelled machines. For an efficient use, these machines require the adequacy of the hedge dimensions. In an hedge orchard at Herdade da Torre das Figueiras, Portugal (39º 04’ N, 07º 29’W), a trial to evaluate the effect of different pruning solutions on the harvesting machine efficiency was made. Treatments were: T0 (farmer pruning) - summer pruning (mechanically from 2018 to 2022), post-harvest pruning (mechanically in 2017 and manually in 2020) and winter pruning (manually in 2018 and mechanically followed by manually in 2017); T1- manual pruning from 2018 until 2022; T2 - post-harvest pruning (mechanically from 2017 until 2021) and winter pruning (manually in 2018, 2020 and 2022) and T3 - summer pruning (mechanically from 2018 until 2022), post-harvest pruning (mechanically in 2017) and winter pruning (manually in 2018, 2020 and 2022). Significant differences were registered between years in the losses of almonds to the ground. Significant differences between treatments were obtained in some years. It is impossible to avoid almond losses to the ground. Pruning interventions based on cuts in canopy faces proved to be the most effective.

The objective of this study was to evaluate the influence of different plant population densities on the mechanized processes of peanut sowing and harvesting, using Statistical Process Control as a tool to monitor and analyze the quality of mechanized operations. The experiment was conducted in a commercial field in Taiúva, São Paulo, Brazil, during the 2024/2025 growing season, using the cultivar IAC 505. Four sowing densities were evaluated: 19, 21, 23, and 25 seeds m⁻¹. The quality indicators assessed were: after sowing (mean emergence time, emergence rate index, and intra-row spacing); before harvesting (cone index and soil water content); and after harvesting (maturity, pod water content, yield, and visible losses). Data were collected sequentially over time and evaluated using Statistical Process Control, allowing the efficiency of mechanized operations to be identified. The densities of 19 and 23 seeds m⁻¹ provided better performance in the mechanized sowing and harvesting operations of peanut. The density of 19 seeds m⁻¹ showed greater stand uniformity, improved intra-row spacing distribution, and more suitable pod conditions at the time of harvesting, resulting in greater efficiency in the crop’s productive potential.

Airblast sprayers for tree crops play a key role in the effectiveness of plant protection treatments in orchards. In particular, the airflow direction generated by radial fans significantly affects droplet transport, canopy penetration, and spray drift. This study investigates the airflow direction produced by a radial fan installed on a commercial airblast sprayer (Frutteto 1000, DSM), used as a reference case. A detailed 3D CAD model of the fan assembly was developed, including blades, deflectors, and the external structure. CFD simulations, performed using Autodesk CFD 2024, were carried out to characterize the airflow field under static operating conditions, with particular attention to velocity magnitude, flow direction, and possible asymmetries. The comparison between numerical predictions and experimental observations showed good agreement in terms of overall flow direction and main aerodynamic patterns. The combined numerical–experimental approach proved to be effective for analyzing airflow behavior in radial fan sprayers and represents a valuable tool for design optimization and improved spray application efficiency.

Fatigue failure in agricultural machinery drivetrain is affected by dynamic loads generated by soil–machine interaction. Under unstructured operating environment, variations in normal load and travel speed continuously change the torque transmitted to components such as motors and reducers. These fluctuating loads directly influence durability and fatigue life. Field experiments are commonly used to evaluate fatigue damage. However, the temporal and spatial variability of soil conditions limits repeatability and quantitative load analysis. Soil bin tests provide a controlled environment in which terramechanics-related variables can be regulated. While previous soil bin studies focused on draft force and tire slip, quantitative evaluation of fatigue damage in agricultural off-road machinery systems remains limited. In this study, soil bin experiments were conducted under normal loads of 0-40 kg and traveling speeds of 0.5-3.0 km/h. Torque and rotational speed of the motor–reducer unit were measured for each condition. The data were analyzed using rainflow counting and Miner’s rule to construct load spectrum and estimate cumulative fatigue damage. The results provide basic data for reliability-based durability design and fatigue life prediction of agricultural power transmission systems, including mechanical and electric axles. The data can also be used to validate virtual engineering-based design models.