Neural networks in closed-loop systems: Verification using interval arithmetic and formal prover

被引:0
作者
Rossi, Federico [1 ]
Bernardeschi, Cinzia [1 ]
Cococcioni, Marco [1 ]
机构
[1] Univ Pisa, Dept Informat Engn, Via G Caruso 16, I-56127 Pisa, PI, Italy
关键词
Cyber-physical systems; Neural networks; Closed-loop control systems; Formal verification; Interval arithmetic;
D O I
10.1016/j.engappai.2024.109238
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Machine Learning approaches have been successfully used for the creation of high-performance control components of cyber-physical systems, where the control dynamics result from the combination of many subsystems. However, these approaches may lack the trustworthiness required to guarantee their reliable application in a safety-critical context. In this paper, we propose a combination of interval arithmetic and theorem-proving verification techniques to analyze safety properties in closed-loop systems that embed neural network components. We show the application of the proposed approach to a model-predictive controller for autonomous driving comparing the neural network verification performance with other existing tools. The results show that open-loop neural network verification through interval arithmetic can outperform existing approaches proving properties with a smaller time overhead. Furthermore, we demonstrate the capability of combining the two approaches to construct a formal model of the network in higher-order logic of the controlled system in a closed-loop.
引用
收藏
页数:11
相关论文
共 50 条
  • [31] Formal verification of cP systems using Coq
    Yezhou Liu
    Radu Nicolescu
    Jing Sun
    Journal of Membrane Computing, 2021, 3 : 205 - 220
  • [32] Formal verification of cP systems using Coq
    Liu, Yezhou
    Nicolescu, Radu
    Sun, Jing
    JOURNAL OF MEMBRANE COMPUTING, 2021, 3 (03) : 205 - 220
  • [33] Using Theorem Provers to Guarantee Closed-Loop System Properties
    Arechiga, Nikos
    Loos, Sarah M.
    Platzer, Andre
    Krogh, Bruce H.
    2012 AMERICAN CONTROL CONFERENCE (ACC), 2012, : 3573 - 3580
  • [34] Closed-loop training of static output feedback neural network controllers for large systems: A distillation case study
    Turan, Evren Mert
    Jaschke, Johannes
    JOURNAL OF PROCESS CONTROL, 2024, 143
  • [35] Learning and Verification of Feedback Control Systems using Feedforward Neural Networks
    Dutta, Souradeep
    Jha, Susmit
    Sankaranarayanan, Sriram
    Tiwari, Ashish
    IFAC PAPERSONLINE, 2018, 51 (16): : 151 - 156
  • [36] Spiking neural P systems: matrix representation and formal verification
    Marian Gheorghe
    Raluca Lefticaru
    Savas Konur
    Ionuţ Mihai Niculescu
    Henry N. Adorna
    Journal of Membrane Computing, 2021, 3 : 133 - 148
  • [37] Fault Diagnosis for a Class of Closed-loop Systems via Deterministic Learning
    Liao, Yuzhe
    Chen, Tianrui
    Wang, Cong
    PROCEEDINGS OF THE 36TH CHINESE CONTROL CONFERENCE (CCC 2017), 2017, : 6992 - 6997
  • [38] Spiking neural P systems: matrix representation and formal verification
    Gheorghe, Marian
    Lefticaru, Raluca
    Konur, Savas
    Niculescu, Ionut Mihai
    Adorna, Henry N.
    JOURNAL OF MEMBRANE COMPUTING, 2021, 3 (02) : 133 - 148
  • [39] Certifying Zeros of Polynomial Systems Using Interval Arithmetic
    Breiding, Paul
    Rose, Kemal
    Timme, Sascha
    ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 2023, 49 (01):
  • [40] Towards Formal Verification of Neural Networks: A Temporal Logic Based Framework
    Wang, Xiaobing
    Yang, Kun
    Wang, Yanmei
    Zhao, Liang
    Shu, Xinfeng
    STRUCTURED OBJECT-ORIENTED FORMAL LANGUAGE AND METHOD (SOFL+MSVL 2019), 2020, 12028 : 73 - 87