PDDLStream: Integrating Symbolic Planners and Blackbox Samplers via Optimistic Adaptive Planning

TL;DR

PDDLStream combines symbolic planning with blackbox sampling techniques using an optimistic adaptive approach, enabling efficient solving of complex robotic planning problems involving high-dimensional continuous variables.

Contribution

It extends PDDL with a declarative interface for blackbox procedures and introduces algorithms that balance exploration and exploitation for faster, more optimized planning solutions.

Findings

Successfully applied to simulated robotic domains.

Achieved faster solutions for tightly-constrained problems.

Produced low-cost solutions through local optimization.

Abstract

Many planning applications involve complex relationships defined on high-dimensional, continuous variables. For example, robotic manipulation requires planning with kinematic, collision, visibility, and motion constraints involving robot configurations, object poses, and robot trajectories. These constraints typically require specialized procedures to sample satisfying values. We extend PDDL to support a generic, declarative specification for these procedures that treats their implementation as black boxes. We provide domain-independent algorithms that reduce PDDLStream problems to a sequence of finite PDDL problems. We also introduce an algorithm that dynamically balances exploring new candidate plans and exploiting existing ones. This enables the algorithm to greedily search the space of parameter bindings to more quickly solve tightly-constrained problems as well as locally optimize to produce low-cost solutions. We evaluate our algorithms on three simulated robotic planning domains as well as several real-world robotic tasks.