# Motion Perception in Reinforcement Learning with Dynamic Objects

arXiv: https://arxiv.org/abs/1901.03162 · HTML: https://mlsanity.com/paper/1901.03162/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Jan 2019.

## Built on by (most cited first, 13 of 13)

- [FlowBot3D: Learning 3D Articulation Flow to Manipulate Articulated Objects](https://mlsanity.com/paper/2205.04382/) (2022)
- [General Flow as Foundation Affordance for Scalable Robot Learning](https://mlsanity.com/paper/2401.11439/) (2024)
- [FabricFlowNet: Bimanual Cloth Manipulation with a Flow-based Policy](https://mlsanity.com/paper/2111.05623/) (2021)
- [Reinforcement Learning with Latent Flow](https://mlsanity.com/paper/2101.01857/) (2021)
- [There Is No Turning Back: A Self-Supervised Approach for Reversibility-Aware Reinforcement Learning](https://mlsanity.com/paper/2106.04480/) (2021)
- [GAFlow: Incorporating Gaussian Attention into Optical Flow](https://mlsanity.com/paper/2309.16217/) (2023)
- [Practical Imitation Learning in the Real World via Task Consistency Loss](https://mlsanity.com/paper/2202.01862/) (2022)
- [ReProHRL: Towards Multi-Goal Navigation in the Real World using Hierarchical Agents](https://mlsanity.com/paper/2308.08737/) (2023)
- [Pre-training of Deep RL Agents for Improved Learning under Domain Randomization](https://mlsanity.com/paper/2104.14386/) (2021)
- [Investigation of Factorized Optical Flows as Mid-Level Representations](https://mlsanity.com/paper/2203.04927/) (2022)
- [Unified Control Framework for Real-Time Interception and Obstacle Avoidance of Fast-Moving Objects with Diffusion Variational Autoencoder](https://mlsanity.com/paper/2209.13628/) (2022)
- [Temporal Shift Reinforcement Learning](https://mlsanity.com/paper/2109.02145/) (2021)
- [Empowering LLMs with Parameterized Skills for Adversarial Long-Horizon Planning](https://mlsanity.com/paper/2509.13127/) (2025)

## Abstract (arXiv)

In dynamic environments, learned controllers are supposed to take motion into account when selecting the action to be taken. However, in existing reinforcement learning works motion is rarely treated explicitly; it is rather assumed that the controller learns the necessary motion representation from temporal stacks of frames implicitly. In this paper, we show that for continuous control tasks learning an explicit representation of motion improves the quality of the learned controller in dynamic scenarios. We demonstrate this on common benchmark tasks (Walker, Swimmer, Hopper), on target reaching and ball catching tasks with simulated robotic arms, and on a dynamic single ball juggling task. Moreover, we find that when equipped with an appropriate network architecture, the agent can, on some tasks, learn motion features also with pure reinforcement learning, without additional supervision. Further we find that using an image difference between the current and the previous frame as an additional input leads to better results than a temporal stack of frames.