Learning from feedback is a common paradigm to acquire information that is hard to specify a priori. In this work, we consider an agent with a known nominal reward model that captures its high-level task objective. Furthermore, the agent operates subject to constraints that are unknown a priori and must be inferred from human interventions. Unlike existing methods, our approach does not rely on full or partial demonstration trajectories or assume a fully reactive human. Instead, we assume access only to sparse interventions, which may in fact be generated proactively by the human, and we only make minimal assumptions about the human. We provide both theoretical bounds on performance and empirical validations of our method. We show that our method enables an agent to learn a constraint set with high accuracy that generalizes well to new environments within a domain, whereas methods that only consider reactive feedback learn an incorrect constraint set that does not generalize well, making constraint violations more likely in new environments.

Although experts carefully specify the high-level decision-making models in autonomous systems, it is infeasible to guarantee safety across every scenario during operation. We therefore propose a safety metareasoning system that optimizes the severity of the system's safety concerns and the interference to the system's task: the system executes in parallel a task process that completes a specified task and safety processes that each address a specified safety concern with a conflict resolver for arbitration. This paper offers a formal definition of a safety metareasoning system, a recommendation algorithm for a safety process, an arbitration algorithm for a conflict resolver, an application of our approach to planetary rover exploration, and a demonstration that our approach is effective in simulation.

Contemporary planning models and methods often rely on constant availability of free state information at each step of execution. However, autonomous systems are increasingly deployed in the open world where state information may be costly or simply unavailable in certain situations. Failing to account for sensor limitations may lead to costly behavior or even catastrophic failure. While the partially observable Markov decision process (POMDP) can be used to model this problem, solving POMDPs is often intractable. We introduce a planning model called a semi-observable Markov decision process (SOMDP) specifically designed for MDPs where state observability may be intermittent. We propose an approach for solving SOMDPs that uses memory states to proactively plan for the potential loss of sensor information while exploiting the unique structure of SOMDPs. Our theoretical analysis and empirical evaluation demonstrate the advantages of SOMDPs relative to existing planning models.

Planning for autonomous operation in unknown environments poses a number of technical challenges. The agent must ensure robustness to unknown phenomena, un-predictable variation in execution, and uncertain resources, all while maximizing its objective. These challenges are ex-acerbated in the context of space missions where uncertainty is often higher, long communication delays necessitate robust autonomous execution, and severely constrained computational resources limit the scope of planning techniques that can be used. We examine this problem in the context of a Europa Lander concept mission where an autonomous lander must collect valuable data and communicate that data back to Earth. We model the problem as a hierarchical task network, framing it as a utility maximization problem constrained by a strictly monotonically decreasing energy resource. We propose a novel deterministic planning framework that uses periodic replanning and sampling-based optimization to better handle model uncertainty and execution variation, while remaining computationally tractable. We demonstrate the efficacy of our framework through simulations of a Europa Lander concept mission in which our approach outperforms several baselines in utility maximization and robustness.

Autonomous systems often use approximate planners that exploit state abstractions to solve large MDPs in real-time decision-making problems. However, these planners can eliminate details needed to produce effective behavior in autonomous systems. We therefore propose a novel model, a partially abstract MDP, with a set of abstract states that each compress a set of ground states to condense irrelevant details and a set of ground states that expand from a set of expanded abstract states to retain relevant details. This papers offers (1) a definition of a partially abstract MDP that (2) generalizes its ground MDP and its abstract MDP and exhibits bounded optimality depending on its abstract MDP along with (3) a lazy algorithm for planning and execution in autonomous systems. The result is a scalable approach that computes near-optimal solutions to large problems in minutes rather than hours.

Despite considerable efforts by human designers, accounting for every unique situation that an autonomous robotic system deployed in the real world could face is often an infeasible task. As a result, many such deployed systems still rely on human assistance in various capacities to complete certain tasks while staying safe. Competence-aware systems (CAS) is a recently proposed model for reducing such reliance on human assistance while in turn optimizing the system’s global autonomous operation by learning its own competence. However, such systems are limited by a fixed model of their environment and may perform poorly if their a priori planning model does not include certain features that emerge as important over the course of the system’s deployment. In this paper, we propose a method for improving the competence of a CAS over time by identifying important state features missing from the system’s model and incorporating them into its state representation, thereby refining its state space. Our approach exploits information that exists in the standard CAS model and adds no extra work to the human. The result is an agent that better predicts human involvement, improving its competence, reliability, and overall performance.