ISAACS: Iterative Soft Adversarial Actor-Critic for Safety

Kai Chieh Hsu, Duy P. Nguyen, Jaime F. Fisac

Research output: Contribution to journalConference articlepeer-review


The deployment of robots in uncontrolled environments requires them to operate robustly under previously unseen scenarios, like irregular terrain and wind conditions. Unfortunately, while rigorous safety frameworks from robust optimal control theory scale poorly to high-dimensional nonlinear dynamics, control policies computed by more tractable “deep” methods lack guarantees and tend to exhibit little robustness to uncertain operating conditions. This work introduces a novel approach enabling scalable synthesis of robust safety-preserving controllers for robotic systems with general nonlinear dynamics subject to bounded modeling error, by combining game-theoretic safety analysis with adversarial reinforcement learning in simulation. Following a soft actor-critic scheme, a safety-seeking fallback policy is co-trained with an adversarial “disturbance” agent that aims to invoke the worst-case realization of model error and training-to-deployment discrepancy allowed by the designer's uncertainty. While the learned control policy does not intrinsically guarantee safety, it is used to construct a real-time safety filter with robust safety guarantees based on forward reachability rollouts. This safety filter can be used in conjunction with a safety-agnostic control policy, precluding any task-driven actions that could result in loss of safety. We evaluate our learning-based safety approach in a 5D race car simulator, compare the learned safety policy to the numerically obtained optimal solution, and empirically validate the robust safety guarantee of our proposed safety filter against worst-case model discrepancy.

Original languageAmerican English
Pages (from-to)90-103
Number of pages14
JournalProceedings of Machine Learning Research
StatePublished - 2023
Externally publishedYes
Event5th Annual Conference on Learning for Dynamics and Control, L4DC 2023 - Philadelphia, United States
Duration: Jun 15 2023Jun 16 2023

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Control and Systems Engineering
  • Statistics and Probability


  • Adversarial Reinforcement Learning
  • Hamilton Jacobi Reachability Analysis
  • Model Predictive Safety Filter

Cite this