Publications

Publications

Injecting Hallucinations in Autonomous Vehicles: A Component-Agnostic Safety Evaluation Framework

Nascimento, A.; Shimanuki, G.; Vismari, L.; Camargo Jr., J.; Almeida Jr., J.; Cugnasca, P.; Queiroz, A.; Bailenson, J.

arXiv. 2025. Under Review

Paper

We introduce a new safety framework that finds critical autonomous vehicle weaknesses by making the system 'hallucinate' fake sensor data.

CORTEX-AVD: A Framework for CORner Case Testing and EXploration in Autonomous Vehicle Development

Shimanuki, G.; Nascimento, A.; Vismari, L.; Camargo Jr., J.; Almeida Jr., J.; Cugnasca, P.

arXiv. 2025. Under Review

Paper

We introduce CORTEX-AVD, an open-source framework that uses Genetic Algorithms to automatically exploit corner cases for AVs from simple text descriptions.

Navigating the Edge with the State-of-the-Art Insights into Corner Case Identification and Generation for Enhanced Autonomous Vehicle Safety

Shimanuki, G.; Nascimento, A.; Vismari, L.; Camargo Jr., J.; Almeida Jr., J.; Cugnasca, P.

arXiv. 2025. Under Review

Paper

This systematic literature review maps 110 selected studies on Autonomous Vehicle corner case generation to identify the current state-of-the-art, key research gaps, and future directions.

Automatic Generation of Critical Scenarios for Evaluating Intelligent Control of Autonomous Vehicles in a Simulated Environment

Shimanuki, G.

Monograph for Qualifying Exam – M.Sc. in Computer Engineering, Escola Politécnica da Universidade de São Paulo (Poli-USP). 2024.

Enhancing Academic Performance Prediction: A Comprehensive Comparison of Machine Learning Techniques and Metrics

Shimanuki, Gabriel; Nascimento, Alexandre; and Queiroz, Anna C. M.

ISLA 2024 Proceedings. 2.. 2024.

Paper

This study compares ten machine learning models and various metrics to find the most accurate and reliable methods for predicting student academic performance.

Making More with Less: Improving Software Testing Outcomes Using a Cross-Project and Cross-Language ML Classifier Based on Cost-Sensitive Training

Nascimento, A. M., Shimanuki, G. K. G., & Dias, L. A. V.

Applied Sciences, 14(11), 4880. 2024.

Paper

We introduce a cost-sensitive, cross-project, and cross-language ML classifier that improves software testing outcomes even with limited resources.

Increasing the Reliability of Autonomous Vehicle Controls Through the Creation of more Robust Traffic Datasets

Shimanuki, G.

Monograph for Senior Thesis Project – B.Sc. in Computer Engineering, Escola Politécnica da Universidade de São Paulo (Poli-USP). 2023.

Poster

An Enhanced Artificial Neural Network Approach To Predict Student Dropout From Imbalanced Datasets

Nascimento, Alexandre; Queiroz, Anna C. M.; and Godoy Shimanuki, Gabriel Kenji

ISLA 2023 Proceedings. 15.. 2023.

Paper

We propose an enhanced Artificial Neural Network approach to more accurately predict student dropout, even when working with highly imbalanced datasets.

Analysis of the Influence of the Safety Zone on the Effectiveness of the Lateral Control Algorithm for Lane Change in Autonomous Vehicles

Shimanuki, G.

30th University of São Paulo's International Symposium of Undergraduate Research. 2022.

Paper Poster

This study uses an AV simulator (OpenDS) to analyze how adding a virtual 'Safety Zone' to a control algorithm improves vehicle safety during lane conversions.

(Originally presented in Portuguese. Short Paper.)

Analysis of the Influence of the Safety Zone on the Effectiveness of the Lane Crossing Control Algorithm for Autonomous Vehicles

Shimanuki, G.

29th University of São Paulo's International Symposium of Undergraduate Research. 2021.

Paper

This simulation study analyzes two 'Safety Zone' strategies, showing that applying the virtual safety buffer to all vehicles (not just the AV) is more effective for safe lane crossing.

(Originally presented in Portuguese. Short Paper.)

Influence of the Safety Zone on Accident Risk with Autonomous Vehicles

Shimanuki, G.

28th University of São Paulo's International Symposium of Undergraduate Research. 2020.

Paper

We use a simulation environment to assess how the size of an AV's virtual 'Safety Zone' impacts collision risk during a lane crossing.

(Originally presented in Portuguese. Short Paper.)

Set of Mock Exams for the ENEM (Brazilian National High School Exam), 1st edition

Nascimento, A., Garcia, W., Garcia, A., Queiroz, A., Garcia, E., Bandeira, F., Shimanuki, G., & Satin, L.

Editora Foco. ISBN: 978-6-55515-820-5. 2023.

Website

This book provides 5 complete, official ENEM exams with deeply-commented answers and performance tracking tools to help students master test-taking strategies.

(Originally printed in Portuguese)

---