Enhancing Tactile-based Reinforcement Learning for Robotic Control

arXiv 2510.21609 ↗First posted 24 Oct 2025cs.RO cs.LGcited by 5

On the map: Deep reinforcement learning methods → RL for robot locomotion and manipulation

Abstract (arXiv)

Achieving safe, reliable real-world robotic manipulation requires agents to evolve beyond vision and incorporate tactile sensing to overcome sensory deficits and reliance on idealised state information. Despite its potential, the efficacy of tactile sensing in reinforcement learning (RL) remains inconsistent. We address this by developing self-supervised learning (SSL) methodologies to more effectively harness tactile observations, focusing on a scalable setup of proprioception and sparse binary contacts. We empirically demonstrate that sparse binary tactile signals are critical for dexterity, particularly for interactions that proprioceptive control errors do not register, such as decoupled robot-object motions. Our agents achieve superhuman dexterity in complex contact tasks (ball bouncing and Baoding ball rotation). Furthermore, we find that decoupling the SSL memory from the on-policy memory can improve performance. We release the Robot Tactile Olympiad (RoTO) benchmark to standardise and promote future research in tactile-based manipulation. Project page: https://elle-miller.github.io/tactile_rl

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods. Builds on 61 of its 77 references that are in the map. First posted on arXiv 24 Oct 2025.

Builds on

61 of 77 references.

Closest ancestors · 2

Same direction, newest first

Other references · 59

2017Proximal Policy Optimization Algorithmscited by 15,215
2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,501
2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 1,653
2020Mastering Atari with Discrete World Modelscited by 1,007
2020CURL: Contrastive Unsupervised Representations for Reinforcement Learningcited by 760
2018Learning Dexterous In-Hand Manipulationcited by 687
2020Learning Quadrupedal Locomotion over Challenging Terraincited by 604
2020Reinforcement Learning with Augmented Datacited by 468
2016Reinforcement Learning with Unsupervised Auxiliary Taskscited by 404
2020Data-Efficient Reinforcement Learning with Self-Predictive Representationscited by 318
2023Orbit: A Unified Simulation Framework for Interactive Robot Learning Environmentscited by 295

38 more

2015Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Imagescited by 274
2019Improving Sample Efficiency in Model-Free Reinforcement Learning from Imagescited by 274
2020Decoupling Representation Learning from Reinforcement Learningcited by 269
2019Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent Variable Modelcited by 231
2019Deep Dynamics Models for Learning Dexterous Manipulationcited by 206
2022DeXtreme: Transfer of Agile In-hand Manipulation from Simulation to Realitycited by 149
2019Unsupervised State Representation Learning in Ataricited by 144
2018State Representation Learning for Control: An Overviewcited by 127
2021Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learningcited by 123
2022In-Hand Object Rotation via Rapid Motor Adaptationcited by 118
2022Visual Dexterity: In-Hand Reorientation of Novel and Complex Object Shapescited by 101
2021The Distracting Control Suite -- A Challenging Benchmark for Reinforcement Learning from Pixelscited by 100
2023General In-Hand Object Rotation with Vision and Touchcited by 99
2019Making Sense of Vision and Touch: Learning Multimodal Representations for Contact-Rich Taskscited by 84
2023Rotating without Seeing: Towards In-hand Dexterity through Touchcited by 84
2021Decoupling Value and Policy for Generalization in Reinforcement Learningcited by 82
2023Dexterity from Touch: Self-Supervised Pre-Training of Tactile Representations with Robotic Playcited by 62
2023Hyperparameters in Reinforcement Learning and How To Tune Themcited by 59
2020Predictive Information Accelerates Learning in RLcited by 56
2023Robot Synesthesia: In-Hand Manipulation with Visuotactile Sensingcited by 56
2022skrl: Modular and Flexible Library for Reinforcement Learningcited by 49
2022Visuo-Tactile Transformers for Manipulationcited by 47
2015From Pixels to Torques: Policy Learning with Deep Dynamical Modelscited by 45
2019Manipulation by Feel: Touch-Based Control with Deep Predictive Modelscited by 36
2023See to Touch: Learning Tactile Dexterity through Visual Incentivescited by 35
2024Multimodal Visual-Tactile Representation Learning through Self-Supervised Contrastive Pre-Trainingcited by 31
2023Neural feels with neural fields: Visuo-tactile perception for in-hand manipulationcited by 28
2023Conditional Mutual Information for Disentangled Representations in Reinforcement Learningcited by 27
2023The Power of the Senses: Generalizable Manipulation from Vision and Touch through Masked Multimodal Learningcited by 26
2025Parkour in the Wild: Learning a General and Extensible Agile Locomotion Policy Using Multi-expert Distillation and RL Fine-tuningcited by 26
2021SwingBot: Learning Physical Features from In-hand Tactile Exploration for Dynamic Swing-up Manipulationcited by 24
2022Investigating the Properties of Neural Network Representations in Reinforcement Learningcited by 24
2024AnyRotate: Gravity-Invariant In-Hand Object Rotation with Sim-to-Real Touchcited by 21
2022Learning Purely Tactile In-Hand Manipulation with a Torque-Controlled Handcited by 20
2022Temporal Disentanglement of Representations for Improved Generalisation in Reinforcement Learningcited by 19
2024RoboPack: Learning Tactile-Informed Dynamics Models for Dense Packingcited by 19
2022Light-weight probing of unsupervised representations for Reinforcement Learningcited by 15
2018Leveraging Contact Forces for Learning to Graspcited by 13

50 most cited of 59 All 59 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 5.

5 papers.

2026Human-Centric Transferable Tactile Pre-Training for Dexterous Robotic Manipulationcited by 5
2026HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodimentscited by 4
2026Egocentric Tactile and Proximity Sensors as Observation Priors for Humanoid Collision Avoidancecited by 1
2026roto 2.0: The Robot Tactile Olympiadcited by 1
2026Beyond Binary: Sim-to-Real Dexterous Manipulation with Physics-Grounded Contact Representationcited by 0

Similar, not cited

Close in wording, no citation either way