A General Safety Framework for Learning-Based Control in Uncertain Robotic Systems
From MaRDI portal
Publication:5223788
Abstract: The proven efficacy of learning-based control schemes strongly motivates their application to robotic systems operating in the physical world. However, guaranteeing correct operation during the learning process is currently an unresolved issue, which is of vital importance in safety-critical systems. We propose a general safety framework based on Hamilton-Jacobi reachability methods that can work in conjunction with an arbitrary learning algorithm. The method exploits approximate knowledge of the system dynamics to guarantee constraint satisfaction while minimally interfering with the learning process. We further introduce a Bayesian mechanism that refines the safety analysis as the system acquires new evidence, reducing initial conservativeness when appropriate while strengthening guarantees through real-time validation. The result is a least-restrictive, safety-preserving control law that intervenes only when (a) the computed safety guarantees require it, or (b) confidence in the computed guarantees decays in light of new observations. We prove theoretical safety guarantees combining probabilistic and worst-case analysis and demonstrate the proposed framework experimentally on a quadrotor vehicle. Even though safety analysis is based on a simple point-mass model, the quadrotor successfully arrives at a suitable controller by policy-gradient reinforcement learning without ever crashing, and safely retracts away from a strong external disturbance introduced during flight.
Recommendations
- Provably safe and robust learning-based model predictive control
- Robust Safety-Critical Control for Dynamic Robotics
- Adaptive learning control of uncertain robotic systems
- Reinforcement learning endowed with safe veto policies to learn the control of linked-multicomponent robotic systems
- scientific article; zbMATH DE number 2117027
- scientific article; zbMATH DE number 18041
- Probabilistic Model Predictive Safety Certification for Learning-Based Control
- A predictive safety filter for learning-based control of constrained nonlinear dynamical systems
- Safe control of nonlinear systems in LPV framework using model-based reinforcement learning
Cited in
(38)- Data-driven fault-tolerant formation control for nonlinear quadrotors under multiple simultaneous actuator faults
- Safe-visor architecture for sandboxing (AI-based) unverified controllers in stochastic cyber-physical systems
- Learning-based symbolic abstractions for nonlinear control systems
- Safe exploration in model-based reinforcement learning using control barrier functions
- Online data-enabled predictive control
- SAMBA: safe model-based \& active reinforcement learning
- Robust learning-based MPC for nonlinear constrained systems
- A predictive safety filter for learning-based control of constrained nonlinear dynamical systems
- Sim-to-lab-to-real: safe reinforcement learning with shielding and generalization guarantees
- Learning for Constrained Optimization: Identifying Optimal Active Constraint Sets
- Robust Control for Dynamical Systems with Non-Gaussian Noise via Formal Abstractions
- Bayesian optimization with safety constraints: safe and automatic parameter tuning in robotics
- Online learning‐based model predictive control with Gaussian process models and stability guarantees
- Online learning constrained model predictive control based on double prediction
- Data‐enabled predictive control for quadcopters
- Safe reinforcement learning: A control barrier function optimization approach
- Adaptive control Lyapunov function based model predictive control for continuous nonlinear systems
- Separation of learning and control for cyber-physical systems
- Strategy synthesis for partially-known switched stochastic systems
- Almost surely safe exploration and exploitation for deep reinforcement learning with state safety estimation
- Verification-guided programmatic controller synthesis
- Safety reinforcement learning control via transfer learning
- Safe adaptive output-feedback optimal control of a class of linear systems
- Combining learning and control in linear systems
- Safe autonomy under perception uncertainty using chance-constrained temporal logic
- Closed-loop stability analysis of deep reinforcement learning controlled systems with experimental validation
- Safe fixed-time reinforcement learning for nonlinear zero-sum games with obstacle avoidance awareness
- Secondary safety control for systems with sector bounded nonlinearities
- All-time safety and sample-efficient meta update for online safe meta reinforcement learning under Markov task transition
- Policy-based primal-dual methods for concave CMDP with variance reduction
- Lagrangian-based online safe reinforcement learning for state-constrained systems
- Faster algorithm and sharper analysis for constrained Markov decision process
- An optimistic approach to cost-aware predictive control
- Safe reinforcement learning for optimal tracking of continuous-time nonlinear systems
- Online Gaussian process learning based adaptive safe actor-critic control for continuous-time nonlinear systems
- Convergence and sample complexity of natural policy gradient primal-dual methods for constrained MDPs
- \texttt{DATA-DRIVEN PRONTO}: a model-free solution for numerical optimal control
- Enforcing almost-sure reachability in POMDPs
This page was built for publication: A General Safety Framework for Learning-Based Control in Uncertain Robotic Systems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5223788)
