Skill Transfer for Temporal Task Specification

Deploying robots in real-world environments, such as households and manufacturing lines, requires generalization across novel task specifications without violating safety constraints. Linear temporal logic (LTL) is a widely used task specification language with a compositional grammar that naturally induces commonalities among tasks while preserving safety guarantees. However, most prior work on reinforcement learning with LTL specifications treats every new task independently, thus requiring large amounts of training data to generalize. We propose LTL-Transfer, a zero-shot transfer algorithm that composes task-agnostic skills learned during training to safely satisfy a wide variety of novel LTL task specifications. Experiments in Minecraft-inspired domains show that after training on only 50 tasks, LTL-Transfer can solve over 90% of 100 challenging unseen tasks and 100% of 300 commonly used novel tasks without violating any safety constraints. We deployed LTL-Transfer at the task-planning level of a quadruped mobile manipulator to demonstrate its zero-shot transfer ability for fetch-and-deliver and navigation tasks.

Authors

• Jason Xinyu Liu
• Ankit Shah
• Eric Rosen
• Mingxi Jia
• George Konidaris 0001
• Stefanie Tellex

Keywords

• Training
• Navigation
• Training data
• Reinforcement learning
• Manipulators
• Safety
• Specification languages
• Specific Tasks
• Transferable Skills
• Temporal Task
• Navigation Task
• Safety Constraints
• Temporal Logic
• Linear Logic
• Mobile Manipulator
• Specific Types
• Training Set
• State Space
• Transfer Learning
• Grid Cells
• Training Tasks
• Termination Condition
• Relaxed State
• Graph Neural Networks
• Policy Learning
• Reinforcement Learning Algorithm
• Test Task
• Outgoing Edges
• Boolean Function
• Simulation Domain
• Task Environment
• Cardinal Directions
• Failure Conditions
• Starting State
• Types Of Tasks
• Environmental Conditions