An official website of the United States government
Offline safe reinforcement learning (OSRL) aims to learn policies with high rewards while satisfying safety constraints solely from data collected offline. However, the learned policies often struggle to handle states and actions that are not present or out-of-distribution (OOD) from the offline dataset, which can result in violation of the safety constraints or overly conservative behaviors during their online deployment. Moreover, many existing methods are unable to learn policies that can adapt to varying constraint thresholds. To address these challenges, we propose constraint-conditioned actor-critic (CCAC), a novel OSRL method that models the relationship between state-action distributions and safety constraints, and leverages this relationship to regularize critics and policy learning. CCAC learns policies that can effectively handle OOD data and adapt to varying constraint thresholds. Empirical evaluations on the benchmarks show that CCAC significantly outperforms existing methods for learning adaptive, safe, and high-reward policies.
more »
« less
Dynamic Model Predictive Shielding for Provably Safe Reinforcement Learning
Among approaches for provably safe reinforcement learning, Model Predictive Shielding (MPS) has proven effective at complex tasks in continuous, high-dimensional state spaces, by leveraging a backup policy to ensure safety when the learned policy attempts to take risky actions. However, while MPS can ensure safety both during and after training, it often hinders task progress due to the conservative and task-oblivious nature of backup policies. This paper introduces Dynamic Model Predictive Shielding (DMPS), which optimizes reinforcement learning objectives while maintaining provable safety. DMPS employs a local planner to dynamically select safe recovery actions that maximize both short-term progress as well as long-term rewards. Crucially, the planner and the neural policy play a synergistic role in DMPS. When planning recovery actions for ensuring safety, the planner utilizes the neural policy to estimate long-term rewards, allowing it to observe beyond its short-term planning horizon. Conversely, the neural policy under training learns from the recovery plans proposed by the planner, converging to policies that are both high-performing and safe in practice. This approach guarantees safety during and after training, with bounded recovery regret that decreases exponentially with planning horizon depth. Experimental results demonstrate that DMPS converges to policies that rarely require shield interventions after training and achieve higher rewards compared to several state-of-the-art baselines
more »
« less
- Award ID(s):
- 2505865
- PAR ID:
- 10631853
- Publisher / Repository:
- Neurips 2024
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This paper presents an integrated motion planning system for autonomous vehicle (AV) parking in the presence of other moving vehicles. The proposed system includes 1) a hybrid environment predictor that predicts the motions of the surrounding vehicles and 2) a strategic motion planner that reacts to the predictions. The hybrid environment predictor performs short-term predictions via an extended Kalman filter and an adaptive observer. It also combines short-term predictions with a driver behavior cost-map to make long-term predictions. The strategic motion planner comprises 1) a model predictive control-based safety controller for trajectory tracking; 2) a search-based retreating planner for finding an evasion path in an emergency; 3) an optimization-based repairing planner for planning a new path when the original path is invalidated. Simulation validation demonstrates the effectiveness of the proposed method in terms of initial planning, motion prediction, safe tracking, retreating in an emergency, and trajectory repairing.more » « less
-
In recent years, Deep Reinforcement Learning (DRL) has emerged as an effective approach to solving real-world tasks. However, despite their successes, DRL-based policies suffer from poor reliability, which limits their deployment in safety-critical domains. Various methods have been put forth to address this issue by providing formal safety guarantees. Two main approaches include shielding and verification. While shielding ensures the safe behavior of the policy by employing an external online component (i.e., a “shield”) that overrides potentially dangerous actions, this approach has a significant computational cost as the shield must be invoked at runtime to validate every decision. On the other hand, verification is an offline process that can identify policies that are unsafe, prior to their deployment, yet, without providing alternative actions when such a policy is deemed unsafe. In this work, we present verification-guided shielding — a novel approach that bridges the DRL reliability gap by integrating these two methods. Our approach combines both formal and probabilistic verification tools to partition the input domain into safe and unsafe regions. In addition, we employ clustering and symbolic representation procedures that compress the unsafe regions into a compact representation. This, in turn, allows to temporarily activate the shield solely in (potentially) unsafe regions, in an efficient manner. Our novel approach allows to significantly reduce runtime overhead while still preserving formal safety guarantees. We extensively evaluate our approach on two benchmarks from the robotic navigation domain, as well as provide an in-depth analysis of its scalability and completeness. 1more » « less
-
null (Ed.)This paper reports on developing an integrated framework for safety-aware informative motion planning suitable for legged robots. The information-gathering planner takes a dense stochastic map of the environment into account, while safety constraints are enforced via Control Barrier Functions (CBFs). The planner is based on the Incrementally-exploring Information Gathering (IIG) algorithm and allows closed-loop kinodynamic node expansion using a Model Predictive Control (MPC) formalism. Robotic exploration and information gathering problems are inherently path-dependent problems. That is, the information collected along a path depends on the state and observation history. As such, motion planning solely based on a modular cost does not lead to suitable plans for exploration. We propose SAFE-IIG, an integrated informative motion planning algorithm that takes into account: 1) a robot’s perceptual field of view via a submodular information function computed over a stochastic map of the environment, 2) a robot’s dynamics and safety constraints via discrete-time CBFs and MPC for closedloop multi-horizon node expansions, and 3) an automatic stopping criterion via setting an information-theoretic planning horizon. The simulation results show that SAFE-IIG can plan a safe and dynamically feasible path while exploring a dense map.more » « less
-
While Deep Reinforcement Learning (DRL) has achieved remarkable success across various domains, it remains vulnerable to occasional catastrophic failures without additional safeguards. An effective solution to prevent these failures is to use a shield that validates and adjusts the agent’s actions to ensure compliance with a provided set of safety specifications. For real-world robotic domains, it is essential to define safety specifications over continuous state and action spaces to accurately account for system dynamics and compute new actions that minimally deviate from the agent’s original decision. In this paper, we present the first shielding approach specifically designed to ensure the satisfaction of safety requirements in continuous state and action spaces, making it suitable for practical robotic applications. Our method builds upon realizability, an essential property that confirms the shield will always be able to generate a safe action for any state in the environment. We formally prove that realizability can be verified for stateful shields, enabling the incorporation of non-Markovian safety requirements, such as loop avoidance. Finally, we demonstrate the effectiveness of our approach in ensuring safety without compromising the policy’s success rate by applying it to a navigation problem and a multi-agent particle environment1. Keywords: Shielding, Reinforcement Learning, Safety, Roboticsmore » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Conference Paper:
- The DOI is not currently available.
