Data collection is one of the most relevant topics of modern automation and industry. It is usually a costly and time-consuming task, especially in continuous processes. Our case study takes place in a sugar cane mill. The required continuous operation of a belt conveyor for bagasse transportation makes it a critical system in the overall production process. Therefore, a predictive maintenance tool is highly applicable here. We identified bearing housings as critical points for data collection intended for prognostics of the conveyor. However, given the number of points, the cost of a commercial solution becomes unfeasible by our company. This paper reports the development of low-cost devices for measurements and wireless transmission of vibration and temperature data from bearing housings. We assessed several sensor options and made decisions based on a cost-suitability commitment, which led to the design of the electronic devices. The devices were tested for correct operation, reliability (99%), and relative measurement errors under 1.2%. From the tests, we conclude that our proposal is appropriate for our case study’s industrial needs and budget restrictions.

Ensuring safe coexistence between human operators and industrial robot manipulators is a critical challenge in collaborative manufacturing environments. Existing approaches rely either on dedicated safety-rated hardware, which is expensive and difficult to retrofit, or on purely vision-based classifiers that discard the precise kinematic state available from the robot controller, leading to unresolved visual ambiguities when different joint configurations produce similar appearances from fixed camera viewpoints. Kinematics-only approaches, while precise, lack the spatial context needed to disambiguate configurations near workspace boundaries. We propose RGBJointsNet, a late-fusion multimodal deep learning classifier that combines RGB visual features extracted by a frozen EfficientNet-B2 convolutional backbone with a compact kinematic stream processing the 12-dimensional joint angle vector of a dual-UR5 robotic cell. The model maps each observation to one of five mutually exclusive workspace zones: rest ($C_0$), nominal ($C_1$), extended ($C_2$), shared/collision-risk ($C_3$), and joint-limit/singularity ($C_4$). A dedicated simulation environment built on ROS 2 Humble Hawksbill and Gazebo Classic 11 was used to generate a labelled dataset of 54,309 frames and 162,927 RGB images from three calibrated overhead cameras, with analytic ground-truth labels derived from closed-form forward kinematics. Training on a CPU with a feature-caching strategy brings the per-epoch wall-clock time to seconds, making the approach tractable without GPU hardware. On the held-out test set, the model achieves 87.1% overall accuracy and a macro-averaged F1 score of 90.0%, with near-perfect recall of 99.3% for the safety-critical shared zone $C_3$. The trained classifier is integrated as an ROS 2 inference node capable of running at 10 $\mathrm{Hz}$ on a standard workstation. Our results demonstrate that joint angle information is a decisive complement to RGB imagery for fine-grained, safety-oriented workspace classification in simulation-derived settings.

Monitoring vegetation-index dynamics in tropical regions remains challenging due to persistent cloud contamination, landscape heterogeneity, and the lack of standardised and reproducible analytical workflows. This paper presents an open-source, fully reproducible end-to-end methodology for multitemporal vegetation monitoring using Sentinel-2 Level-2A (L2A) Bottom-of-Atmosphere (BOA) reflectance imagery. The methodology was applied to Military Grid Reference System (MGRS) tile T18NVL in the Colombian Eje Cafetero region (4.43°N–5.43°N, 74.91°W–75.90°W) for the 2017–2025 period. The proposed workflow integrates storage-efficient direct extraction of spectral reflectance from compressed Standard Archive Format for Europe (SAFE) archives using the Geospatial Data Abstraction Library (GDAL) /vsizip/ interface, per-pixel cloud and shadow masking based on the Sentinel-2 Scene Classification Layer (SCL), computation of Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Soil-Adjusted Vegetation Index (SAVI), and Normalized Difference Moisture Index (NDMI) spectral indices, a diagnostic Random Forest experiment based on threshold-labelled spectral classes, non-parametric Mann–Kendall trend analysis with Sen’s slope estimation, and external agreement assessment against European Space Agency (ESA) WorldCover 10 m 2020 and Google Earth Pro reference data. The methodology reduced per-scene I/O time by approximately 98% without additional disk overhead while retaining a median of 57.9% valid pixels under a mean scene cloud fraction of 35.4%. Mann–Kendall analysis detected no statistically significant long-term trend in any of the four vegetation indices. Seasonal NDVI peaks during September–November were consistent with the bimodal regional precipitation regime, supporting temporal coherence in the satellite-derived vegetation-index response. External agreement was low, with an Overall Accuracy (OA) of 20.2% against ESA WorldCover and 19.9% against Google Earth Pro, indicating systematic over-prediction of woody and mixed-canopy vegetation classes. These results show that single-date optical spectral indices are insufficient for reliable thematic separation of shade-grown coffee, secondary forest, and dense forest within heterogeneous tropical landscapes. The complete version-controlled codebase is publicly available to support methodological reproducibility and adaptation across data-scarce tropical regions.

Simulation-based reinforcement-learning locomotion depends on the physical fidelity of the underlying model. This work presents a subsystem-level modelling, validation, and control framework for a 12-DoF biped robot, combining Denavit–Hartenberg kinematics, Euler–Lagrange dynamics, a Discrete Euler–Lagrange reference integrator, Hunt–Crossley contact, and Soft Actor-Critic training in PyBullet. Validation is scoped. For the fixed-hip leg, numerical damped-least-squares inverse kinematics achieved a round-trip error of 0.017 ± 0.022 mm, while a gravity path-integral test produced a residual of 0.006 J. On a one-DoF reference problem, DEL bounded energy error under a coarse-step stress test, whereas at the 1 ms training step, RK4 was more accurate; no RL-scale DEL advantage was established. Contact realism remained inconclusive because the available prescribed-penetration analysis was not a dynamically consistent whole-body impact test. The same nominal parameters were used in PyBullet for locomotion training, without establishing numerical equivalence between the two simulators. Across three asymmetric-reward runs, forward walking dominated final evaluations, but sustained velocity ranged from 0.62 to 1.24 m/s under unequal training budgets. An exploratory hybrid architecture reached 2.38 m/s in one run without controlled ablation. These results demonstrate subsystem-level diagnostics while identifying full-body validation, contact calibration, equal-budget replication, and architectural ablation as necessary future work.

Reinforcement-learning (RL) locomotion results are only as trustworthy as the physical fidelity of the models behind them. This paper presents an end-to-end framework for a 12-degree-of-freedom (DoF) biped robot that couples (i) a complete nonlinear dynamic model derived via Denavit–Hartenberg parameterisation and the Euler–Lagrange formulation with full Christoffel-symbol coupling; (ii) a structure-preserving reference simulator based on Discrete Euler–Lagrange (DEL) variational integration and a Hunt–Crossley viscoelastic contact law; and (iii) a Soft Actor-Critic (SAC) locomotion controller with three-phase curriculum learning. This paper proposes a four-metric validation framework to assess the physical fidelity of the analytical model and reference simulator before any reinforcement learning policy is trained. The model and simulator are validated through four quantitative metrics tied to hardware-measurable references: (M1) forward-inverse kinematics round-trip error of 0.017±0.021 mm (kinematic consistency); (M2) a path-integral gravity residual of 0.006 J (dynamic consistency); (M3) bounded mechanical-energy oscillation with no secular trend under a deliberately coarse-step stress test, in which the DEL integrator reduces secular drift by three orders of magnitude relative to fourth-order Runge-Kutta at 44% lower cost (energy conservation); and (M4) a smooth heel-strike force profile consistent with body-weight-normalised force-plate data (contact realism). These four metrics provide an auditable, falsifiable, and reproducible procedure to verify that the model loaded into the training environment is physically meaningful before any learning occurs. The validated model parameters are exported to a standard physics engine (PyBullet), in which the SAC policy is trained for 3 million control steps. An asymmetric directional reward eliminates a backward-walking local optimum, raising the sustained mean forward velocity to +1.24 m/s, a 55% improvement over the 0.80 m/s hardware gait of the same-morphology ARON reference design, with a transient peak of 2.796 m/s and 99% of final-phase evaluations walking forward. A subsequent bimodal architecture, a frozen Exp. 4 encoder fused with a trainable 1D-CNN branch over an eight-step observation history, achieves a further 92% improvement, reaching a sustained mean of +2.38 m/s in only 1 M training steps. The framework separates, and makes auditable, the roles of model certification and policy training in simulation-based biped design.