Filters








3,879 Hits in 7.1 sec

Uncertainty Quantification with Statistical Guarantees in End-to-End Autonomous Driving Control [article]

Rhiannon Michelmore, Matthew Wicker, Luca Laurenti, Luca Cardelli, Yarin Gal, Marta Kwiatkowska
2019 arXiv   pre-print
In addition to computing pointwise uncertainty measures that can be computed in real time and with statistical guarantees, we also provide a method for estimating the probability that, given a scenario  ...  In this paper, we develop a framework based on a state-of-the-art simulator for evaluating end-to-end Bayesian controllers.  ...  In this paper we develop a novel framework for evaluating the safety of autonomous driving using end-to-end BNN controllers, that is, controllers in which the end-to-end process, from sensors to actuation  ... 
arXiv:1909.09884v1 fatcat:kpykayzgl5gd7gmmsdrgu7ldtq

Marginally calibrated response distributions for end-to-end learning in autonomous driving [article]

Clara Hoffmann, Nadja Klein
2021 arXiv   pre-print
These learners must provide reliable uncertainty estimates for their predictions in order to meet safety requirements and initiate a switch to manual control in areas of high uncertainty.  ...  End-to-end learners for autonomous driving are deep neural networks that predict the instantaneous steering angle directly from images of the ahead-lying street.  ...  An end-to-end learner trained on this data could be directly employed on real highways. We use it to provide a realistic quantification of uncertainty of end-to-end learners for highway driving.  ... 
arXiv:2110.01050v1 fatcat:iolauf4mirbyxhypclb4tdadjy

Planning and Decision-Making for Autonomous Vehicles

Wilko Schwarting, Javier Alonso-Mora, Daniela Rus
2018 Annual Review of Control Robotics and Autonomous Systems  
Recent advances in the field of perception, planning, and decision-making for autonomous vehicles have led to great improvements in functional capabilities, with several prototypes already driving on our  ...  In this review, we provide an overview of emerging trends and challenges in the field of intelligent and autonomous, or self-driving, vehicles.  ...  We briefly introduce methods for parallel autonomy, where a human is still in control of the vehicle, and then focus on autonomous vehicles.  ... 
doi:10.1146/annurev-control-060117-105157 fatcat:hgrhw76idbbdrct742bbhnsqem

Improving Trustworthiness of AI Disease Severity Rating in Medical Imaging with Ordinal Conformal Prediction Sets [article]

Charles Lu, Anastasios N. Angelopoulos, Stuart Pomerantz
2022 arXiv   pre-print
A lack of statistically rigorous uncertainty quantification is a significant factor undermining trust in AI results.  ...  Recent developments in distribution-free uncertainty quantification present practical solutions for these issues by providing reliability guarantees for black-box models on arbitrary data distributions  ...  driving control [22] .  ... 
arXiv:2207.02238v1 fatcat:ez2bsw6l2bgc7lfnywqq6pccme

Robustness in Cyber-Physical Systems (Dagstuhl Seminar 16362)

Martin Fränzle, James Kapinski, Pavithra Prabhakhar, Marc Herbstritt
2017 Dagstuhl Reports  
At the same time, these systems are increasing in complexity at an alarming rate, making it difficult to produce system designs with guaranteed robust performance.  ...  Electronically controlled systems have become pervasive in modern society and are increasingly being used to control safety-critical applications, such as medical devices and transportation systems.  ...  timing (in the sense of worst-case execution times of tasks, worstcase end-to-end latencies in circuits or reactive systems, etc.).  ... 
doi:10.4230/dagrep.6.9.29 dblp:journals/dagstuhl-reports/FranzleKP16 fatcat:uiivaekedzdqdpmn2shsecrpve

Mind the Gap! A Study on the Transferability of Virtual vs Physical-world Testing of Autonomous Driving Systems [article]

Andrea Stocco, Brian Pulfer, Paolo Tonella
2021 arXiv   pre-print
Safe deployment of self-driving cars (SDC) necessitates thorough simulated and in-field testing.  ...  Most testing techniques consider virtualized SDCs within a simulation environment, whereas less effort has been directed towards assessing whether such techniques transfer to and are effective with a physical  ...  Kwiatkowska, “Uncertainty Quantification with Statistical ary Computation Conference, GECCO ’21, (New York, NY, USA), Guarantees in End-to-End Autonomous Driving Control,” in 2020  ... 
arXiv:2112.11255v1 fatcat:aia44y5s2vamnabgnofo3lxo2a

Safe AI – How is this Possible? [article]

Harald Rueß, Simon Burton
2022 arXiv   pre-print
Ttraditional safety engineering is coming to a turning point moving from deterministic, non-evolving systems operating in well-defined contexts to increasingly autonomous and learning-enabled AI systems  ...  We outline some of underlying challenges of safe AI and suggest a rigorous engineering framework for minimizing uncertainty, thereby increasing confidence, up to tolerable levels, in the safe behavior  ...  Uncertainty due to open-ended operating contexts and safe control thereof is dramatically reduced in current automotive practice, by collecting all kinds of possible driving scenarios by means of global  ... 
arXiv:2201.10436v2 fatcat:lu5ibn3qc5hormd4w6zjmszplq

Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing

Johannes Betz, Hongrui Zheng, Alexander Liniger, Ugo Rosolia, Phillip Karle, Madhur Behl, Venkat Krovi, Rahul Mangharam
2022 IEEE Open Journal of Intelligent Transportation Systems  
We focus on the field of autonomous racecars only and display the algorithms, methods, and approaches used in the areas of perception, planning, control, and end-to-end learning.  ...  The rising popularity of self-driving cars has led to the emergence of a new research field in recent years: Autonomous racing.  ...  ." 8 which contributed to the creation of Section V in this survey paper.  ... 
doi:10.1109/ojits.2022.3181510 fatcat:2uu6aowqrrewfbkfrsrmqxpyei

Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing [article]

Johannes Betz, Hongrui Zheng, Alexander Liniger, Ugo Rosolia, Phillip Karle, Madhur Behl, Venkat Krovi, Rahul Mangharam
2022 arXiv   pre-print
We focus on the field of autonomous racecars only and display the algorithms, methods and approaches that are used in the fields of perception, planning and control as well as end-to-end learning.  ...  The rising popularity of self-driving cars has led to the emergence of a new research field in the recent years: Autonomous racing.  ...  Classic autonomous driving software pipeline in comparison to partial and full end-to-end software pipeline. Fig. 11 . 11 Fig. 11.  ... 
arXiv:2202.07008v1 fatcat:hwhp43thevd2bighs7y7j7qnam

A Formal Safety Characterization of Advanced Driver Assist Systems in the Car-Following Regime with Scenario-Sampling [article]

Bowen Weng, Minghao Zhu, Keith Redmill
2022 arXiv   pre-print
In this paper, we propose a guaranteed unbiased and sampling efficient scenario-based safety evaluation framework inspired by the previous work on ϵδ-almost safe set quantification.  ...  The capability to follow a lead-vehicle and avoid rear-end collisions is one of the most important functionalities for human drivers and various Advanced Driver Assist Systems (ADAS).  ...  Furthermore, the action u is typically determined by a certain feedback control policy u = π(s, ω s ; ω u ), (4) with s, ω s the same with what we have defined above, and the uncertainties ω u ∈ W u .  ... 
arXiv:2202.08935v2 fatcat:wtjcwd7torh7ppz575vedwgxce

Self-driving car safety quantification via component-level analysis [article]

Juozas Vaicenavicius and Tilo Wiklund and Austė Grigaitė and Antanas Kalkauskas and Ignas Vysniauskas and Steven Keen
2021 arXiv   pre-print
In this paper, we present a rigorous modular statistical approach for arguing safety or its insufficiency of an autonomous vehicle through a concrete illustrative example.  ...  A simple concrete automated braking example studied illustrates how separate perception system and operational design domain statistical analyses can be used to prove or disprove safety at the vehicle  ...  Acknowledgements We would like to thank the anonymous reviewers for constructive comments. Our thanks go also to Mark Costin, PhD, at NVIDIA for useful discussions and practical feedback.  ... 
arXiv:2009.01119v4 fatcat:uvff254mh5ennd73a2i7s7oywm

How to Certify Machine Learning Based Safety-critical Systems? A Systematic Literature Review [article]

Florian Tambon, Gabriel Laberge, Le An, Amin Nikanjam, Paulina Stevia Nouwou Mindom, Yann Pequignot, Foutse Khomh, Giulio Antoniol, Ettore Merlo, François Laviolette
2021 arXiv   pre-print
In total, we identified 217 papers covering topics considered to be the main pillars of ML certification: Robustness, Uncertainty, Explainability, Verification, Safe Reinforcement Learning, and Direct  ...  However, including it in so-called 'safety-critical' systems such as automotive or aeronautic has proven to be very challenging, since the shift in paradigm that ML brings completely changes traditional  ...  Acknowledgements We would like to thank the following authors (in no particular order) who kindly provided us feedback about our review of their work: Mahum Naseer, Hoang-Dung Tran, Jie Ren, David Isele  ... 
arXiv:2107.12045v3 fatcat:43vqxywawbeflhs6ehzovvsevm

Use of Hybrid Causal Logic Method for Preliminary Hazard Analysis of Maritime Autonomous Surface Ships

Di Zhang, Zhepeng Han, Kai Zhang, Jinfen Zhang, Mingyang Zhang, Fan Zhang
2022 Journal of Marine Science and Engineering  
Finally, the accident probability of autonomy level III MASS is calculated in practice through historical data and a test ship with both an autonomous and a remote navigation mode in Wuhan and Nanjing,  ...  Furthermore, the fault tree (FT) method is utilized to analyze mechanical events in ESD.  ...  As for human-and organization-related events, due to their uncertainties, we applied the Bayesian Belief Network (BBN) to analyze in detail the influence factors based on the experimental statistics This  ... 
doi:10.3390/jmse10060725 fatcat:mzaqjdoslzaf3f7j5w4yolvx2y

A testing framework for predictive driving features with an electronic Horizon

M. Elgharbawy, A. Schwarzhaupt, R. Arenskrieger, H. Elsayed, M. Frey, F. Gauterin
2019 Transportation Research Part F: Traffic Psychology and Behaviour  
Even if the widespread use of full automation is not imminent, the vision of accident-free driving accelerates the further development of driver assistance functions to autonomous vehicle stages on the  ...  The status quo evaluation refers to large-scale verification as one of the decisive challenges for the economical, reliable and safe use of automated driving functions in truck series development.  ...  Uncertainty quantification can provide information that is employed in object plausibility within sensor fusion algorithms. Two types of uncertainties can be distinguished.  ... 
doi:10.1016/j.trf.2017.08.002 fatcat:ebqind2uejfnjaxq36aq7nuxja

Security and Privacy Dimensions in Next Generation DDDAS/Infosymbiotic Systems: A Position Paper

Li Xiong, Vaidy Sunderam
2015 Procedia Computer Science  
While every single smartphone or wearable device is potentially a sensor with powerful computing and data capabilities, privacy and security in the context of human participants must be addressed to leverage  ...  We propose a security and privacy preserving framework for next generation systems that harness the full power of the DDDAS paradigm while (1) ensuring provable privacy guarantees for sensitive data; (  ...  at higher levels, which in turn drive the collection process.  ... 
doi:10.1016/j.procs.2015.05.357 fatcat:yvddllnnfbbclaun4wevqtbmzm
« Previous Showing results 1 — 15 out of 3,879 results