Formal verification of ACAS X, an industrial airborne collision avoidance system

Jeannin, Jean-Baptiste; Ghorbal, Khalil; Kouskoulas, Yanni; Gardner, Ryan W.; Schmidt, Aurora; Zawadzki, Erik; Platzer, André

doi:10.1109/emsoft.2015.7318268

Cited by 25 publications

(15 citation statements)

References 13 publications

(32 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Datasets. We use MNIST [22] and ACAS Xu [12,17] as the datasets in our experiments. MNIST contains 60 000 grayscale handwritten digits of the size 28 × 28.…”

Section: Experimental Evaluationmentioning

confidence: 99%

Improving Neural Network Verification through Spurious Region Guided Refinement

Yang

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We propose a spurious region guided refinement approach for robustness verification of deep neural networks. Our method starts with applying the DeepPoly abstract domain to analyze the network. If the robustness property cannot be verified, the result is inconclusive. Due to the over-approximation, the computed region in the abstraction may be spurious in the sense that it does not contain any true counterexample. Our goal is to identify such spurious regions and use them to guide the abstraction refinement. The core idea is to make use of the obtained constraints of the abstraction to infer new bounds for the neurons. This is achieved by linear programming techniques. With the new bounds, we iteratively apply DeepPoly, aiming to eliminate spurious regions. We have implemented our approach in a prototypical tool DeepSRGR. Experimental results show that a large amount of regions can be identified as spurious, and as a result, the precision of DeepPoly can be significantly improved. As a side contribution, we show that our approach can be applied to verify quantitative robustness properties.

show abstract

“…Datasets. We use MNIST [22] and ACAS Xu [12,17] as the datasets in our experiments. MNIST contains 60 000 grayscale handwritten digits of the size 28 × 28.…”

Section: Experimental Evaluationmentioning

confidence: 99%

Improving Neural Network Verification through Spurious Region Guided Refinement

Yang

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…36 However, there are a few challenging case studies that have been developed through scaling formal methods approaches. In Jeannin et al, 37 the authors present a formal verification of ACAS X, an airborne collision avoidance system, using a hybrid systems theorem prover. In Klein et al, 11 the authors present the formal, machine-checked verification of the seL4 microkernel.…”

Section: Related Workmentioning

confidence: 99%

A formal approach to rigorous development of critical systems

Singh

Lawford

Maibaum

et al. 2021

J Software Evolu Process

View full text Add to dashboard Cite

Safety critical systems, such as medical, automotive, and avionics systems, play an important role in our daily lives. Increasing demand for new technologies in these safety critical systems requires rapid adoption of commercial hardware and software. However, the adoption of new hardware and software increases life‐threatening vulnerabilities. To aid in the reduction of these vulnerabilities and system failures, this paper proposes a framework based on formal methods for developing safety‐critical systems from requirements analysis to code generation. This framework includes a development process for documenting system requirements using tabular expressions, automatic formal model generation from the documented requirements, verification and validation of the generated formal models using proof techniques and animations, interactive simulation for validating the required behavior of the developed models by enabling domain experts to observe the system states according to, and finally, code generation from the formal model into a desired language. A prototype toolchain is developed to automate this framework. An assessment of the proposed framework is undertaken through a case study: insulin infusion pump (IIP).

show abstract

“…An interesting side effect of digitalization is the arrival of companies on the software market that so far had no major stake in software. Here is an opportunity for formal methods and formal verification, in particular, since software verification tools are as well applicable to cyber-physical systems [52,73] (see Challenge C.5).…”

Section: Industrial and Societal Contextmentioning

confidence: 99%

Deductive Software Verification: From Pen-and-Paper Proofs to Industrial Tools

Hähnle

Huisman

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Deductive software verification aims at formally verifying that all possible behaviors of a given program satisfy formally defined, possibly complex properties, where the verification process is based on logical inference. We follow the trajectory of the field from its inception in the late 1960s via its current state to its promises for the future, from pen-and-paper proofs for programs written in small, idealized languages to highly automated proofs of complex library or system code written in mainstream languages. We take stock of the state-of-art and give a list of the most important challenges for the further development of the field of deductive software verification.

show abstract

Formal verification of ACAS X, an industrial airborne collision avoidance system

Cited by 25 publications

References 13 publications

Improving Neural Network Verification through Spurious Region Guided Refinement

Improving Neural Network Verification through Spurious Region Guided Refinement

A formal approach to rigorous development of critical systems

Deductive Software Verification: From Pen-and-Paper Proofs to Industrial Tools

Contact Info

Product

Resources

About