Learning safe neural network controllers with barrier certificates

Authors:
Open Access
Publisher website
Google Scholar
Abstract
We provide a new approach to synthesize controllers for nonlinear continuous dynamical systems with control against safety properties. The controllers are based on neural networks (NNs). To certify the safety property we utilize barrier functions, which are represented by NNs as well. We train the controller-NN and barrier-NN simultaneously, achieving a verification-in-the-loop synthesis. We provide a prototype tool nncontroller with a number of case studies. The experiment results confirm the feasibility and efficacy of our approach.
Funding Information
  • National Natural Science Foundation of China (61702425)
  • National Natural Science Foundation of China (61972385)
  • National Natural Science Foundation of China (61902325)
  • Fundamental Research Funds for the Central Universities (SWU117058)
  • National Natural Science Foundation of China (61872340)
  • Guangdong Science and Technology Department Grant (2018B010107004)
  • Overseas Grant of the State Key Laboratory of Novel Software Technology (KFKT2018A16)
  • Natural Science Foundation of Guangdong Province of China (2019A1515011689)
  • National Natural Science Foundation of China (61672435)
  • National Natural Science Foundation of China (61732019)
  • Capacity Development Grant of Southwest University (SWU116007)
  • National Natural Science Foundation of China (62032019)

This publication has 37 references indexed in Scilit: