zur Hauptnavigation springen zum Inhaltsbereich springen

BayWISS-Kolleg Mobilität & Verkehr www.baywiss.de

Projekte im Kolleg Mobilität und Verkehr

© chuttersnap / unsplash.com'

Validation Machine Learning-based Highly Automated Driving Functions by Diversity

Recently, the trend to implement Machine Learning (ML) algorithms in various engineering disciplines has signi_cantly increased, e.g., in computer vision, robotics, or biomedical engineering. A key enabler of these ML algorithms is the vast abundance of data (e.g., picture, video, sensor data, etc.) and the increase in computational power (e.g., the use of GPUs), which can be used to train complex ML models. Despite the ability of these ML algorithms to model highly complex functions and perform classi_cation tasks with high accuracy, the lack of validation methods for ML algorithms remains a hindering factor in their implementation in safety related functions, e.g., in highly automated driving functions.

On the one hand, the di_culty to explain the classi_cation decisions of ML algorithms makes validation di_cult, as these algorithms are mostly seen as black-box methods. In general, training ML algorithms is non-deterministic, which creates the necessity to obtain high quality training data to ensure that the ML algorithm learns meaningful representations and generalises correctly. Moreover, since ML methods purely rely on data, the existing physical models are ignored. To this end, the abundance of data-based and model-based methods enable engineers to fuse the outputs of independently optimised algorithms to create con_dence intervals or plausibility envelopes around the outputs of ML algorithms. With these quality metrics and plausibility claims of ML algorithms, we aim to create methods which can be used to validate safety related functions in highly automated driving.

MITGLIED IM KOLLEG

von bis

Verbundkolleg Mobilität & Verkehr

Dr. Oliver de Candido

Oliver de Candido

Technische Universität München

Publikationen und Poster

 

J. Y. Tee, O. De Candido, W. Utschick, and P. Geiger, 09/2023, On Learning the Tail Quantiles of Driving Behavior Distributions via Quantile Regression and Flows, 26th IEEE International Conference on Intelligent Transportation Systems ITSC 2023; Bilbao, Bizkaia, Spain

M. Henneberg, C. Eghtebas, O. De Candido, K. Kunze, and J. A. Ward, 04/2023, Detecting an Offset-Adjusted Similarity Score based on Duchenne Smiles, CHI EA '23: Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems; Hamburg, Germany, 10.1145/3544549.3585709

K. Klein, O. De Candido, and W. Utschick, 07/2023, Interpretable Classifiers based on Time-Series Motifs for Lane Change Prediction, IEEE Transactions on Intelligent Vehicles, Volume: 8, Issue: 7, 10.1109/TIV.2023.3276650

Oliver De Candido, Michael Koller, Wolfgang Utschick, 04/2022, Encouraging Validatable Features in Machine Learning-based Highly Automated Driving Functions, IEEE Transactions on Intelligent Vehicles, DOI:10.1109/TIV.2022.3171215

Oliver De Candido, Xinyang Li, Wolfgang Utschick, 06/2022, An Analysis of Distributional Shifts in Automated Driving Functions in Highway Scenarios, Helsinki, Finland, DOI:10.1109/VTC2022-Spring54318.2022.9860453

Philipp Joppich, Sebastian Dorn, Oliver De Candido, Jakob Knollmüller, Wolfgang Utschick, 07/2022, Classification and Uncertainty Quantification of Corrupted Data Using Supervised Autoencoders, Paris, France, 41st International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, DOI:10.3390/psf2022005012

Tobias Uelwer, Felix Michels, Oliver De Candido, 12/2022, Evaluating Robust Perceptual Losses for Image Reconstruction, New Orleans, USA, NeurIPS 2022, I Can't Believe It's Not Better Workshop: Understanding Deep Learning Through Empirical Falsification

De Candido, O.; Binder, M.; Utschick, W., 07/2021, An Interpretable Lane Change Detector Algorithm based on Deep Autoencoder Anomaly Detection, 2021 IEEE Intelligent Vehicles Symposium (IV), Nagoya, Japan (Onlinekonferenz), DOI: 10.1109/IV48863.2021.9575599, peer-reviewed

Gallitz O., De Candido O., Botsch M., Utschick W., 09/2021, Interpretable Early Prediction of Lane Changes Using a Constrained Neural Network Architecture, 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), Indianapolis, IN, USA (Onlinekonferenz), DOI: 10.1109/itsc48978.2021.9564555, peer-reviewed

Joppich, Philipp; Dorn, Sebastian; De Candido, Oliver; Utschick, Wolfgang; Knollmüller, Jakob, 05/2021, Classification and Uncertainty Quantification of Corrupted Data using Semi-Supervised Autoencoders, arXiv preprint arXiv:2105.13393

Konstantinidis Fabian; Hofmann Ulrich; Sackmann Moritz; Thielecke Jorn; De Candido Oliver; Utschick Wolfgang, 09/2021, Parameter Sharing Reinforcement Learning for Modeling Multi-Agent Driving Behavior in Roundabout Scenarios, 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), Indianapolis, IN, USA (Onlinekonferenz), DOI: 10.1109/itsc48978.2021.9565031, peer-reviewed

Uelwer Tobias; Michels Felix; De Candido Oliver, 09/2021, Learning to Detect Adversarial Examples Based on Class Scores, KI 2021: Advances in Artificial Intelligence, Onlinekonferenz, DOI: 10.1007/978-3-030-87626-5_17, peer-reviewed

De Candido, O., Gallitz, O., Melz, R., Botsch, M., Utschick,W., 10/2020, Interpretable Machine Learning Structure for an Early Prediction of Lane Changes, Onlinekonferenz: Artificial Neural Networks and Machine Learning – ICANN 2020. DOI: 10.1007/978-3-030-61609-0_27, peer-reviewed

De Candido, O., Koller, M., Gallitz, O., Melz, R., Botsch, M., Utschick,W., 2020, Towards Feature Validation in Time to Lane Change Classification using Deep Neural Networks. Sep. 20-23, 2020, The 23rd IEEE Intelligent Transportation Systems Conference (ITSC). DOI: 10.1109/ITSC45102.2020.9294555, peer-reviewed

Gallitz, O., De Candido, O., Melz, R., Botsch, M., Utschick,W., 2020, Interpretable Machine Learning Structure for an Early Prediction of Lane Changes.ICANN (1): 337-349

K. Klein, O. De Candido, and W. Utschick, 07/2023, Interpretable Classifiers based on Time-Series Motifs for Lane Change Prediction, IEEE Transactions on Intelligent Vehicles, Volume: 8, Issue: 7, 10.1109/TIV.2023.3276650

Koordination des Verbundkollegs Mobilität und Verkehr

Treten Sie mit uns in Kontakt. Wir freuen uns auf Ihre Fragen und Anregungen zum Verbundkolleg Mobilität und Verkehr.

Dr. Monika Kolpatzik

Dr. Monika Kolpatzik

Koordinatorin BayWISS-Verbundkolleg Mobilität & Verkehr

Technische Hochschule Ingolstadt
Doctoral School
Esplanade 10
85049 Ingolstadt

Telefon: +49 841 93481560
mobilitaet-verkehr.vk [ at ] baywiss.de

Marina Schleicher

Marina Schleicher

Koordinatorin BayWISS-Verbundkolleg Mobilität & Verkehr

Technische Hochschule Ingolstadt
Doctoral School
Esplanade 10
85049 Ingolstadt

Telefon: +49 841 93483539
mobilitaet-verkehr.vk [ at ] baywiss.de