Task-space guidance improves robot learning for tabletop manipulation
Task-Space Imitation Guidance for Efficient Reinforcement Learning
Robotics
Summary
Robots often struggle to learn tasks when rewards for their actions are very rare or unclear. The authors propose a new method called TIGER that uses imitation to estimate short-term progress towards a goal. Instead of directly copying actions, TIGER guides the robot with continuous hints about how well it is moving towards the task, making learning faster and safer. Tests in simulations and real robots show better early learning and fewer safety issues compared to older methods.
What this means in practice
- •For industrial robotics teams: Improve early training speed and safety of robots performing precise tabletop assembly tasks using imitation-based progress signals.
- •For automation engineers: Develop safer robot deployment by reducing unsafe actions during initial learning in sparse-reward manipulation tasks.
Authors
Salar Asayesh, Hossein Darani, Todd Cao, Evgeny Andriash, Mani Ranjbar
Abstract
We introduce Task-Space Imitation Guidance for Efficient Reinforcement Learning (TIGER), a reward-construction and pretraining framework for sparse-reward tabletop robotic manipulation. TIGER treats an action-chunked imitation policy not as an executable controller or action prior, but as a local task-space progress estimator: predicted action chunks are converted, using controller-aware action-to-motion mapping, into short-horizon end-effector references, and the RL agent receives dense progress rewards toward these references while the sparse environment reward remains the dominant objective. During pretraining, TIGER uses imitation-guided look-ahead signals to relax conservative value penalties for actions predicted to make task-space progress, reducing off-manifold exploration during early online RL. Across simulation and real-robot experiments, TIGER improves early sample efficiency and reduces measured safety violations while matching or improving final success rates relative to prior RL and IL-RL baselines on the evaluated tasks.