Carlos M.A. Diogo

Dr. Carlos M.A. Diogo

IN+ C. for Innovation Technology & Policy Research, Universidade de Lisboa - Instituto Superior Tecnico, Lisbon, Lisboa, Portugal
orcid

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
18 January 2023 in Progress in Aerospace Sciences
Verified
Journal Article
Closed access
FAFrederico Afonso
Frederico Afonso
MSMartin Sohst
Martin Sohst
CDCarlos M.A. Diogo
Carlos M.A. Diogo
SRSimão Rodrigues
Simão Rodrigues
AFAna Ferreira
Ana Ferreira
ASAfzal Suleman
Afzal Suleman

15 August 2024 in Drones
Verified
Journal Article
Open access
PDF file
CDCarlos M.A. Diogo
Carlos M.A. Diogo
Edgar FernandesEdgar Fernandes
Edgar Fernandes
Fixed-Wing Unmanned Aerial Vehicles (UAVs) have been improving significantly in application and versatility, sharing design similarities with airplanes, particularly at the design stage, when the take-off mass is used to estimate other characteristics. In this work, an internal database of UAVs is built to allow their comparison with airplanes under different parameters and assess key differences in patterns across UAV powertrains. The existing literature on speed vs. take-off mass is updated with 534 UAV entries, and a range vs. take-off mass diagram is created with 503 UAVs and 193 airplanes. Additionally, different transportation efficiency metrics are compared between UAVs and airplanes, highlighting scenarios advantageous for UAVs. A new paradigm focused on useful energy is then used to understand the underlying effectiveness of UAV implementations. Increasing useful energy is more effective in increasing the speed, transport work, and surveying work of internal combustion UAVs and more effective in increasing the range and endurance of battery-electric UAVs. Finally, it is observed that the mass of all fixed-wing aerial vehicles, both UAVs and airplanes, except for battery electric and solar, adheres to a well-defined scaling law based on useful energy. A parallel to this scaling law is suggested to describe future battery-electric UAVs and airplanes.

1 February 2018 in Finite Elements in Analysis and Design
Verified
Journal Article
Closed access
CDCarlos M.A. Diogo
Carlos M.A. Diogo
Diogo NetoDiogo Neto
Diogo Neto
Marta OliveiraMarta Oliveira
Marta Oliveira
JAJ.L. Alves
J.L. Alves
Luis Filipe MenezesLuis Filipe Menezes
Luis Filipe Menezes

31 August 2016 in Journal of Physics: Conference Series
Verified
Conference Paper
Open access
PDF file
Diogo NetoDiogo Neto
Diogo Neto
CDCarlos M.A. Diogo
Carlos M.A. Diogo
TNT F Neves
T F Neves
Marta OliveiraMarta Oliveira
Marta Oliveira
JAJ L Alves
J L Alves
Luis Filipe MenezesLuis Filipe Menezes
Luis Filipe Menezes
Most of sheet metal forming processes comprise intermediate trimming operations to remove superfluous material. These operations are required for subsequent forming operations. On the other hand, the springback is strongly influenced by the trimming operations that change the part stiffness and the stress field. From the numerical point of view, this involves the geometrical trimming of the finite element mesh and subsequent remapping of the state variables. This study presents a remapping method based on Dual Kriging interpolation, specifically developed for hexahedral finite elements, which has been implemented in DD3TRIM in-house code. Its performance is compared with the one of the Incremental Volumetric Remapping method, using the split-ring test to highlight their advantages and limitations. The numerical simulation of the forming processes is performed with DD3IMP finite element solver.

23 April 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MFMargarida Tânger de Oliveira Figueiredo
Margarida Tânger de Oliveira Figueiredo
CDCarlos M.A. Diogo
Carlos M.A. Diogo
Gustavo André PaneiroGustavo André Paneiro
Gustavo André Paneiro
PAPedro Amaral
Pedro Amaral
ACAntónio Alves de Campos
António Alves de Campos
The marble industry relies on proprietary commercial names rather than objective visual categories, creating market inefficiencies for stakeholders who select stones based on appearance. Supervised classification perpetuates this problem by replicating inconsistent commercial labels instead of discovering intrinsic visual structure. We propose an unsupervised pipeline combining a two-stage training strategy: A pure self-supervised pretraining followed by cluster-aware fine-tuning of a DINO Vision Transformer, with empirically selected dimensionality reduction and agglomerative hierarchical clustering. Systematic ablation studies on 1480 marble images spanning 10 commercial varieties validate each design choice: cluster-aware training at k = 10 yields geometrically improved embeddings over the self-supervised baseline (mean Silhouette Score 0.693 ± 0.053 vs. 0.660 ± 0.030; mean Davies–Bouldin Index 0.386 ± 0.075 vs. 0.569 ± 0.012; N = 9 independent evaluations across 3 data partitions × 3 training initializations). The resulting taxonomy reveals three phenomena invisible to commercial classification: cross-category merging of visually indistinguishable stones carrying different market names, intra-category splitting of heterogeneous sub-populations within single varieties, and coherent grouping where commercial and visual boundaries coincide, with all three confirmed in every independent run. We further demonstrate that standard extrinsic metrics are misaligned with unsupervised taxonomy objectives when reference labels encode the inconsistencies the method aims to resolve. Validating this methodology across diverse stone types, larger datasets, and varied acquisition conditions represents a natural and necessary next step toward establishing its cross-domain generalizability.

10 February 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
ACAntónio Alves de Campos
António Alves de Campos
MFMargarida Tânger de Oliveira Figueiredo
Margarida Tânger de Oliveira Figueiredo
CDCarlos M.A. Diogo
Carlos M.A. Diogo
Gustavo André PaneiroGustavo André Paneiro
Gustavo André Paneiro
PAPedro Amaral
Pedro Amaral
Generative Adversarial Networks (GANs) have demonstrated remarkable capabilities for synthesizing photorealistic textures, yet deploying conditional GANs (cGANs) in industrial settings faces two barriers: the prohibitive cost of annotating proprietary data and the uncertain alignment between automated metrics and human perception. This study addresses both challenges for marble texture synthesis. We adapt an unsupervised segmentation pipeline combining Simple Linear Iterative Clustering (SLIC) superpixels, Gaussian Mixture Models (GMMs), and Graph Cut optimization to extract vein structures from 289 industrial scans without manual annotation. We then benchmark four cGAN architectures, a baseline cGAN, Pix2Pix, BicycleGAN, and GauGAN, using a dual-evaluation protocol contrasting automated assessment via pixel-based metrics, structural metrics, statistical metrics, and learned distributional metrics with human-centered assessment. Results reveal a significant metric–perception discrepancy: Pix2Pix achieved the best FID yet received the lowest human ratings due to checkerboard artifacts, whereas GauGAN produced textures statistically indistinguishable from real marble (Visual Turing Pass Rate, VTPR: 0.533; Mean Opinion Score on Marble Authenticity, MOS-MA: 2.89) despite inferior FID (87.3). These findings establish three contributions: (1) an unsupervised annotation-free segmentation pipeline, (2) empirical evidence that automated metrics alone are insufficient for architecture selection, and (3) a dual-evaluation framework validating human-in-the-loop assessment as essential for industrial deployment.