Constrained Belief-Dependent POMDP Planning

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 and safety-related, it is required to introduce a belief-dependent constraint. In this research project we present a novel formulation and for a risk-averse belief-dependent probabilistically constrained continuous POMDP. We investigate different aspects of this framework, and, particularly, introduce adaptive simplification in a probabilistically constrained setting.

Related Publications:

Journal Articles

  1. A. Zhitnikov and V. Indelman, “Simplified Continuous High Dimensional Belief Space Planning with Adaptive Probabilistic Belief-dependent Constraints,” IEEE Transactions on Robotics (T-RO), 2024.
    Zhitnikov24tro.pdf Zhitnikov24tro.slides DOI: 10.1109/TRO.2023.3341625

Theses

  1. A. Zhitnikov, “Simplification for Efficient Decision Making Under Uncertainty with General Distributions,” PhD thesis, Technion - Israel Institute of Technology, 2024.
    Zhitnikov24thesis.pdf Zhitnikov24thesis.slides Zhitnikov24thesis.video

Technical Reports

  1. A. Zhitnikov and V. Indelman, “Anytime Probabilistically Constrained Provably Convergent Online Belief Space Planning,” 2024.
    arXiv: https://arxiv.org/abs/2411.06711
  2. A. Zhitnikov and V. Indelman, “Risk Aware Belief-dependent Constrained Simplified POMDP Planning,” Sep. 2022.
    arXiv: https://arxiv.org/pdf/2209.02679.pdf