Simplified Risk-aware Decision Making with Belief-dependent Rewards in Partially Observable Domains

Zhitnikov, Andrey; Indelman, Vadim

doi:10.1016/j.artint.2022.103775

Cited by 3 publications

(7 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…b t (x t ) P(x t |h t ). In this paper the belief converts the history to a more convenient form, b t and can be used interchangeably with h t , as opposed to our previous work [15].…”

Section: Background and Notationsmentioning

confidence: 99%

“…In this problem the utility conforms to (5). Importantly, as in our previous paper [23] δ and are supplied by the user.…”

Section: Problem Formulation and Approachmentioning

confidence: 99%

“…Recent methods, merging both worlds, build upon the simplification paradigm [14], [15], [16]. These simplificationbased methods finally relax limiting assumptions, e.g., Gaussian belief, Piecewise linearity, or Lipshitz continuity of the reward, and permitted universal belief-dependent rewards such as differential entropy of general beliefs.…”

Section: Introductionmentioning

confidence: 99%

“…A general belief-dependent constraint, however, has not received proper attention so far except in our previous work [23], where we focused on safety and not information gathering tasks.…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, we continue to investigate the facets of our proposed earlier framework [23] of general belief-dependent constrained continuous POMDP. Motivated by information gathering, also called informative planning tasks, we focus on the cumulative form of the constraint in contrast to the multiplicative form as in our previous paper.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Simplified Continuous High Dimensional Belief Space Planning with Adaptive Probabilistic Belief-dependent Constraints

Zhitnikov¹,

Indelman²

2023

Preprint

View full text Add to dashboard Cite

Online decision making under uncertainty in partially observable domains, also known as Belief Space Planning, is a fundamental problem in robotics and Artificial Intelligence. Due to an abundance of plausible future unravelings, calculating an optimal course of action inflicts an enormous computational burden on the agent. Moreover, in many scenarios, e.g., information gathering, it is required to introduce a beliefdependent constraint. Prompted by this demand, in this paper, we consider a recently introduced probabilistic belief-dependent constrained POMDP. We present a technique to adaptively accept or discard a candidate action sequence with respect to a probabilistic belief-dependent constraint, before expanding a complete set of future observations samples and without any loss in accuracy. Moreover, using our proposed framework, we contribute an adaptive method to find a maximal feasible return (e.g., information gain) in terms of Value at Risk for the candidate action sequence with substantial acceleration. On top of that, we introduce an adaptive simplification technique for a probabilistically constrained setting. Such an approach provably returns an identical-quality solution while dramatically accelerating online decision making. Our universal framework applies to any beliefdependent constrained continuous POMDP with parametric beliefs, as well as nonparametric beliefs represented by particles. In the context of an information-theoretic constraint, our presented framework stochastically quantifies if a cumulative information gain along the planning horizon is sufficiently significant (e.g. for, information gathering, active SLAM). We apply our method to active SLAM, a highly challenging problem of high dimensional Belief Space Planning. Extensive realistic simulations corroborate the superiority of our proposed ideas.

show abstract

“…b t (x t ) P(x t |h t ). In this paper the belief converts the history to a more convenient form, b t and can be used interchangeably with h t , as opposed to our previous work [15].…”

Section: Background and Notationsmentioning

confidence: 99%

“…In this problem the utility conforms to (5). Importantly, as in our previous paper [23] δ and are supplied by the user.…”

Section: Problem Formulation and Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…A general belief-dependent constraint, however, has not received proper attention so far except in our previous work [23], where we focused on safety and not information gathering tasks.…”

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations