Enhancing Safety and Formal Verification in Autonomous Systems

Advances in Autonomous Systems and Formal Verification

The recent advancements in the field of autonomous systems have seen a significant shift towards enhancing safety, robustness, and formal verification techniques. Researchers are increasingly focusing on integrating formal methods with machine learning to ensure the reliability and safety of autonomous systems, particularly in critical applications such as autonomous driving and industrial robotics. The field is moving towards more sophisticated hybrid models that combine neural networks with traditional formal verification methods, aiming to leverage the strengths of both approaches. Additionally, there is a growing emphasis on user-friendly tools and frameworks that facilitate the development and validation of these systems, making them more accessible for educational and industrial use.

Noteworthy papers include:

An extension to arbitration graphs for safer autonomous decision-making, validated in autonomous driving scenarios.
A neural network hybrid modeling framework for dynamics learning, promoting interpretability and computational efficiency.
A novel symbolic solver for geometry, Newclid, which is more user-friendly and expands the scope of solvable problems.

Sources

Better Safe Than Sorry: Enhancing Arbitration Graphs for Safe and Robust Autonomous Decision-Making

Efficient Neural Hybrid System Learning and Transition System Abstraction for Dynamical Systems

Beyond object identification: How train drivers evaluate the risk of collision

Newclid: A User-Friendly Replacement for AlphaGeometry

Predicting Lemmas in Generalization of IC3

Proceedings Sixth International Workshop on Formal Methods for Autonomous Systems

Breadboarding the European Moon Rover System: discussion and results of the analogue field test campaign

Grand Challenges in the Verification of Autonomous Systems

Creating a Formally Verified Neural Network for Autonomous Navigation: An Experience Report

Autonomous System Safety Properties with Multi-Machine Hybrid Event-B

Formal Simulation and Visualisation of Hybrid Programs

Model Checking and Verification of Synchronisation Properties of Cobot Welding

Cross--layer Formal Verification of Robotic Systems

Using Formal Models, Safety Shields and Certified Control to Validate AI-Based Train Systems

Model Checking for Reinforcement Learning in Autonomous Driving: One Can Do More Than You Think!