Eric Aislan Antonelo

Universidade Federal de Santa Catarina · Brazil

Unclaimed · AI-structured profileORCID OpenAlex

Autonomous driving and learning-based control researcher at Universidade Federal de Santa Catarina, Brazil

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

Not yet scored: the output axis would measure publication volume and productivity, such as total works and their distribution over time. This snapshot lacks the underlying paper-level output data, so no score or estimate is given.

Impact—

Not yet scored: the impact axis would capture citation-based influence (citation volume and quality relative to field). The snapshot reports h-index 16 and i10-index 22 but lacks the paper-level citation evidence required, so no score is assigned.

Focus0/100

The focus axis scores 0, reflecting an unusually wide topical footprint: OpenAlex topic clusters span five groups — Neural Networks and Reservoir Computing (28 works), Reinforcement Learning in Robotics (23), Neural Networks and Applications (21), Model Reduction and Neural Networks (11), and Robotics and Sensor-Based Localization (10) — while recent field weight splits evenly (0.5/0.5) between two autonomous-driving-related fields.

Collaboration28/100

Scoring 28, the collaboration axis reflects a deliberately compact network: only 3 distinct co-authors in the recent three-year window, with Gustavo Claudio Karl Couto, Gabriel George Zipperer, and Felipe Carlos dos Santos recurring across the cataloged 2026 papers. Antonelo appears as a middle author on the two driving-platform papers and as first and corresponding author on the diffusion-control study.

Activity89/100

The highest radar axis (89) matches the raw trajectory: 20 works in the recent three-year window, and per-year OpenAlex counts climb to 8 (2025) and 7 (2026, to date) — the strongest consecutive years in a record running back to 2005 (3 works), ahead of the previous peak of 6 in 2021.

Reproducibility78/100

Scored 78 from Evolve's reproducibility review of 2 papers (mean 78). The low-cost miniature Ackermann platform and the conditional-diffusion control study rate high reproducibility readiness — the former as an explicitly open, low-cost sim-to-real setup — while the closed-loop BEV evaluation paper rates medium, with verdicts flagging small run counts (5 episodes per town; n=1–3 runs).

Research summary

Eric Aislan Antonelo is a Brazil-based researcher at Universidade Federal de Santa Catarina whose recent work centers on end-to-end autonomous driving and learning-based control. His OpenAlex-derived record (h-index 16, i10-index 22) reaches back to 2005, with topic clusters led by Neural Networks and Reservoir Computing (28 works), Reinforcement Learning in Robotics (23), and Neural Networks and Applications (21); recent field weight splits evenly between autonomous driving / robot learning and robotics / machine learning for end-to-end driving. Activity is the standout axis: 20 works in the most recent three-year window, with per-year counts peaking at 8 in 2025 and 7 in 2026. Three 2026 arXiv papers anchor the Evolve catalog: a closed-loop CARLA evaluation of cross-view-transformer bird's-eye-view maps for behavior cloning; a low-cost, open miniature Ackermann platform with a Webots digital twin for sim-to-real driving; and a first- and corresponding-author study of conditional diffusion models for open-loop control under dry friction and stiction. Evolve scores two of the three at high reproducibility readiness (mean 78 across 2 scored papers). Collaboration is compact — 3 distinct recent co-authors, with Couto, Zipperer, and dos Santos recurring — and review verdicts consistently flag thin experimental statistics, so results read best as promising early evidence.

Collaboration pattern

The recent record shows a compact, stable collaboration pattern: only 3 distinct co-authors in the last three years, with Gustavo Claudio Karl Couto, Gabriel George Zipperer, and Felipe Carlos dos Santos recurring across all three cataloged 2026 papers. Antonelo takes first and corresponding authorship on the diffusion-control study and middle-author positions on the two autonomous-driving platform papers — consistent with a small core team sharing experimental infrastructure (CARLA, Webots) rather than a wide external network.

Research keywords

end-to-end autonomous drivingbehavior cloningbird's-eye view (BEV)cross-view transformerclosed-loop evaluationCARLA simulatorimitation learningsim-to-real transferdiffusion modelslearning-based controlreinforcement learning in roboticsreservoir computing

Highlights

Infraction-free closed-loop driving from predicted BEV maps
In closed-loop CARLA evaluation, KDE-weighted cross-view-transformer bird's-eye-view inputs yielded the only infraction-free episode among behavior-cloning agents using predicted BEV maps — suggestive, though Evolve notes only 5 episodes per town and no variance reporting. Reproducibility readiness: medium.
An open, low-cost miniature platform for sim-to-real driving
The open miniature Ackermann platform pairs a Webots digital twin with sim-augmented behavior cloning; the policy completed all four physical routes, though Evolve flags that physical evidence rests on n=1–3 runs. Reproducibility readiness: high.
Conditional diffusion for open-loop control under stiction
As first and corresponding author, this study shows conditional action-sequence diffusion models reduce terminal error and stuck steps on a 1D dry-friction and stiction benchmark using classifier-free guidance — credible but narrow in scope with thin statistics. Reproducibility readiness: high.

Timeline

2026First- and corresponding-author study of conditional diffusion models for open-loop control under dry friction and stiction.
2026Releases a low-cost, open end-to-end autonomous-driving platform on a miniature Ackermann vehicle with a Webots digital twin.
2026Publishes a closed-loop CARLA evaluation of cross-view-transformer BEV maps as behavior-cloning policy inputs.

Papers in the Evolve catalog

2026-09-05
Closed-Loop Evaluation of Bird's-Eye-View Maps from Cross-View Transformers as Inputs to Behavior-Cloning Policies
2026-09-03
A Low-Cost, Open Platform for End-to-End Autonomous Driving on a Miniature Ackermann Vehicle
2026-09-01first authorcorresponding
A Study of Conditional Diffusion Models for Open-Loop Control under Dry Friction and Stiction

Frequent co-authors

Gustavo Claudio Karl Couto×2 Gabriel George Zipperer×1 Felipe Carlos dos Santos×1

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

Bibliometric indicators (h-index 16, i10-index 22, per-year counts, topic clusters) come from an OpenAlex-derived snapshot; citation-level impact data is not part of this profile.
All three reviewed papers are 2026 arXiv preprints; peer-review status and final publication venues are not reflected here.
Identity is resolved from public OpenAlex bibliographic data; disambiguation relies on that record's clustering and may merge or split same-name authors.
Longer-term topic clusters (e.g., reservoir computing) are inferred from OpenAlex topic assignments only; no paper list outside the three reviewed works was available.