Closed-Loop ACAS Xu Neural Network Verification

Sanaz Sheikhi; Stanley Bak

doi:10.29007/vf8z

Download PDF Open PDF in browser

Closed-Loop ACAS Xu Neural Network Verification

8 pages•Published: October 18, 2023

Sanaz Sheikhi and Stanley Bak

Abstract

Benchmark Proposal: Neural Network Control Systems (NNCS) play critical roles in autonomy. However, verifying their correctness is a substantial challenge. In this paper, we consider the neural network compression of ACAS Xu, a popular benchmark usually considered for open-loop neural network verification. ACAS Xu is an air-to-air collision avoidance system for unmanned aircraft issuing horizontal turn advisories to avoid collision with an intruder aircraft. We propose specific properties and different system assumptions to use this system as a closed-loop NNCS benchmark. We present experimental results for our properties based on randomly generated test cases and provide simulation code.

Keyphrases: ACAS Xu, autonomous systems, dynamical systems, nonlinear systems, test case, verification

In: Goran Frehse and Matthias Althoff (editors). Proceedings of 10th International Workshop on Applied Verification of Continuous and Hybrid Systems (ARCH23), vol 96, pages 1--8

Links:	https://easychair.org/publications/paper/mmSLW
	https://doi.org/10.29007/vf8z

BibTeX entry

@inproceedings{ARCH23:Closed_Loop_ACAS_Xu_Neural,
  author    = {Sanaz Sheikhi and Stanley Bak},
  title     = {Closed-Loop ACAS Xu Neural Network Verification},
  booktitle = {Proceedings of 10th International Workshop on Applied Verification of Continuous and Hybrid Systems (ARCH23)},
  editor    = {Goran Frehse and Matthias Althoff},
  series    = {EPiC Series in Computing},
  volume    = {96},
  pages     = {1--8},
  year      = {2023},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2398-7340},
  url       = {https://easychair.org/publications/paper/mmSLW},
  doi       = {10.29007/vf8z}}

Download PDF Open PDF in browser