ALOHA Unleashed: A Simple Recipe for Robot Dexterity

2024/10/17 by Tony Z. Zhao, Jonathan Tompson, Zhao, Tony Z. +11 · 48 citations
Computer Science · #FOS: Computer and information sciences #Real-Time Systems Scheduling #Robotics (cs.RO)

paper · pdf · doi:10.48550/arxiv.2410.13126

openalex publication_date 2024/10/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Recent work has shown promising results for learning end-to-end robot policies using imitation learning. In this work we address the question of how far can we push imitation learning for challenging dexterous manipulation tasks. We show that a simple recipe of large scale data collection on the ALOHA 2 platform, combined with expressive models such as Diffusion Policies, can be effective in learning challenging bimanual manipulation tasks involving deformable objects and complex contact rich dynamics. We demonstrate our recipe on 5 challenging real-world and 3 simulated tasks and demonstrate improved performance over state-of-the-art baselines. The project website and videos can be found at aloha-unleashed.github.io.

Cited by

• Decompose and Reorganize: Planning with Primitives and Visuomotor Policies Learned from Demonstrations
• RoboCade: Gamifying Robot Data Collection
• Posterior Behavioral Cloning: Pretraining BC Policies for Efficient RL Finetuning
• Evaluating Gemini Robotics Policies in a Veo World Simulator
• ESPADA: Execution Speedup via Semantics Aware Demonstration Data Downsampling for Imitation Learning
• Training-Time Action Conditioning for Efficient Real-Time Chunking
• Hybrid-Diffusion Models: Combining Open-loop Routines with Visuomotor Diffusion Policies
• SpeedAug: Policy Acceleration via Tempo-Enriched Policy and RL Fine-Tuning
• Contact-Rich Robotic Assembly in Construction via Diffusion Policy Learning
• In-N-On: Scaling Egocentric Manipulation with in-the-wild and on-task Data
• VIRAL: Visual Sim-to-Real at Scale for Humanoid Loco-Manipulation
• From Power to Precision: Learning Fine-grained Dexterity for Multi-fingered Robotic Hands
• Learning a Thousand Tasks in a Day
• SeFA-Policy: Fast and Accurate Visuomotor Policy Learning with Selective Flow Alignment
• EveryDayVLA: A Vision-Language-Action Model for Affordable Robotic Manipulation
• Embodiment Transfer Learning for Vision-Language-Action Models
• Tri-Manual Visuomotor Imitation Learning of Robot Policies
• Using Non-Expert Data to Robustify Imitation Learning via Offline Reinforcement Learning
• Botany-Bot: Digital Twin Monitoring of Occluded and Underleaf Plant Structures with Gaussian Splats
• Spatial Forcing: Implicit Spatial Representation Alignment for Vision-language-action Model
• Reinforcement Fine-Tuning of Flow-Matching Policies for Vision-Language-Action Models
• Failure Prediction at Runtime for Generative Robot Policies
• R2RGEN: Real-to-Real 3D Data Generation for Spatially Generalized Manipulation
• Vision-Language-Action Models for Robotics: A Review Towards Real-World Applications
• RTFF: Random-to-Target Fabric Flattening Policy using Dual-Arm Manipulator
• Annotation-Free One-Shot Imitation Learning for Multi-Step Manipulation Tasks
• Hybrid Diffusion for Simultaneous Symbolic and Continuous Planning
• VLBiMan: Vision-Language Anchored One-Shot Demonstration Enables Generalizable Bimanual Robotic Manipulation
• ROPA: Synthetic Robot Pose Generation for RGB-D Bimanual Data Augmentation
• MV-UMI: A Scalable Multi-View Interface for Cross-Embodiment Learning
• Do You Need Proprioceptive States in Visuomotor Policies?
• Residual Off-Policy RL for Finetuning Behavior Cloning Policies
• FILIC: Dual-Loop Force-Guided Imitation Learning with Impedance Torque Control for Contact-Rich Manipulation Tasks
• HDMI: Learning Interactive Humanoid Whole-Body Control from Human Videos
• Right-Side-Out: Learning Zero-Shot Sim-to-Real Garment Reversal
• Self-Improving Embodied Foundation Models
• M4Diffuser: Multi-View Diffusion Policy with Manipulability-Aware Control for Robust Mobile Manipulation
• Learning to Pick: A Visuomotor Policy for Clustered Strawberry Picking
• RaC: Robot Learning for Long-Horizon Tasks by Scaling Recovery and Correction
• A Dataset and Benchmark for Robotic Cloth Unfolding Grasp Selection: The ICRA 2024 Cloth Competition
• Learning in ImaginationLand: Omnidirectional Policies through 3D Generative Models (OP-Gen)
• Reactive In-Air Clothing Manipulation with Confidence-Aware Dense Correspondence and Visuotactile Affordance
• Data Retrieval with Importance Weights for Few-Shot Imitation Learning
• Learning Fast, Tool aware Collision Avoidance for Collaborative Robots
• Exploiting Policy Idling for Dexterous Manipulation
• Constraint-Preserving Data Generation for Visuomotor Policy Learning
• COLLAGE: Adaptive Fusion-based Retrieval for Augmented Policy Learning
• Improving Generalization Ability of Robotic Imitation Learning by Resolving Causal Confusion in Observations

Related

• Dexterous Manipulation through Imitation Learning: A Survey
• Dexterous Imitation Made Easy: A Learning-Based Framework for Efficient Dexterous Manipulation
• DexMimicGen: Automated Data Generation for Bimanual Dexterous Manipulation via Imitation Learning
• A System for Imitation Learning of Contact-Rich Bimanual Manipulation Policies
• H-RDT: Human Manipulation Enhanced Bimanual Robotic Manipulation