# Active Perception and Control from Temporal Logic Specifications

arXiv: https://arxiv.org/abs/1905.03662 · HTML: https://mlsanity.com/paper/1905.03662/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 May 2019.

## Built on by (most cited first, 4 of 4)

- [Joint Learning of Reward Machines and Policies in Environments with Partially Known Semantics](https://mlsanity.com/paper/2204.11833/) (2022)
- [Noisy Symbolic Abstractions for Deep RL: A case study with Reward Machines](https://mlsanity.com/paper/2211.10902/) (2022)
- [CogSense: A Cognitively Inspired Framework for Perception Adaptation](https://mlsanity.com/paper/2107.10456/) (2021)
- [Transfer Learning for Control Systems via Neural Simulation Relations](https://mlsanity.com/paper/2412.01783/) (2024)

## Abstract (arXiv)

Next-generation autonomous systems must execute complex tasks in uncertain environments. Active perception, where an autonomous agent selects actions to increase knowledge about the environment, has gained traction in recent years for motion planning under uncertainty. One prominent approach is planning in the belief space. However, most belief-space planning starts with a known reward function, which can be difficult to specify for complex tasks. On the other hand, symbolic control methods automatically synthesize controllers to achieve logical specifications, but often do not deal well with uncertainty. In this work, we propose a framework for scalable task and motion planning in uncertain environments that combines the best of belief-space planning and symbolic control. Specifically, we provide a counterexample-guided-inductive-synthesis algorithm for probabilistic temporal logic over reals (PRTL) specifications in the belief space. Our method automatically generates actions that improve confidence in a belief when necessary, thus using active perception to satisfy PRTL specifications.