Discovering adversarial driving maneuvers against autonomous vehicles
Over 33% of vehicles sold in 2021 had integrated autonomous driving (AD) systems. While
many adversarial machine learning attacks have been studied against these systems, they …
many adversarial machine learning attacks have been studied against these systems, they …
Risk-aware scene sampling for dynamic assurance of autonomous systems
Autonomous Cyber-Physical Systems must often operate under uncertainties like sensor
degradation and shifts in the operating conditions, which increases its operational risk …
degradation and shifts in the operating conditions, which increases its operational risk …
A scenario-based platform for testing autonomous vehicle behavior prediction models in simulation
Behavior prediction remains one of the most challenging tasks in the autonomous vehicle
(AV) software stack. Forecasting the future trajectories of nearby agents plays a critical role …
(AV) software stack. Forecasting the future trajectories of nearby agents plays a critical role …
Hierarchical potential-based reward shaping from task specifications
The automatic synthesis of policies for robotic-control tasks through reinforcement learning
relies on a reward signal that simultaneously captures many possibly conflicting …
relies on a reward signal that simultaneously captures many possibly conflicting …
Compositional Inductive Invariant Based Verification of Neural Network Controlled Systems
Y Zhou, S Tripakis - NASA Formal Methods Symposium, 2024 - Springer
The integration of neural networks into safety-critical systems has shown great potential in
recent years. However, the challenge of effectively verifying the safety of Neural Network …
recent years. However, the challenge of effectively verifying the safety of Neural Network …
Finding Safety Violations of AI-Enabled Control Systems through the Lens of Synthesized Proxy Programs
Given the increasing adoption of modern AI-enabled control systems, ensuring their safety
and reliability has become a critical task in software testing. One prevalent approach to …
and reliability has become a critical task in software testing. One prevalent approach to …
Boundary adherence and exploration in high dimensions for validation of black-box systems
State space exploration is limited to a testing budget in which a finite amount of resources
can be expended to attain a degree of confidence in the system. A potential approach to …
can be expended to attain a degree of confidence in the system. A potential approach to …
Dynamic, Multi-objective Specification and Falsification of Autonomous CPS
KKC Chang, K Xu, E Kim… - … Conference on Runtime …, 2024 - Springer
Simulation-based falsification has proved to be an effective verification method for cyber-
physical systems. Traditional approaches to falsification take as input a single or a set of …
physical systems. Traditional approaches to falsification take as input a single or a set of …
Querying Labeled Time Series Data with Scenario Programs
D Shanker - arXiv preprint arXiv:2406.17627, 2024 - arxiv.org
In order to ensure autonomous vehicles are safe for on-road deployment, simulation-based
testing has become an integral complement to on-road testing. The rise in simulation testing …
testing has become an integral complement to on-road testing. The rise in simulation testing …
Safe policy improvement in constrained Markov decision processes
L Berducci, R Grosu - … Symposium on Leveraging Applications of Formal …, 2022 - Springer
The automatic synthesis of a policy through reinforcement learning (RL) from a given set of
formal requirements depends on the construction of a reward signal and consists of the …
formal requirements depends on the construction of a reward signal and consists of the …