Policy-Guided Lazy Search with Feedback for Task and Motion Planning

TL;DR

This paper introduces LAZY, a PDDLStream solver that integrates lazy sampling and learned models to efficiently solve complex task and motion planning problems with continuous actions.

Contribution

LAZY is a novel integrated search approach that combines lazy sampling with learned goal-directed policies for improved efficiency in TAMP.

Findings

LAZY significantly reduces runtime compared to existing solvers.

Incorporating learned models improves planning speed and success rate.

The approach generalizes well across various unseen environments.

Abstract

PDDLStream solvers have recently emerged as viable solutions for Task and Motion Planning (TAMP) problems, extending PDDL to problems with continuous action spaces. Prior work has shown how PDDLStream problems can be reduced to a sequence of PDDL planning problems, which can then be solved using off-the-shelf planners. However, this approach can suffer from long runtimes. In this paper we propose LAZY, a solver for PDDLStream problems that maintains a single integrated search over action skeletons, which gets progressively more geometrically informed, as samples of possible motions are lazily drawn during motion planning. We explore how learned models of goal-directed policies and current motion sampling data can be incorporated in LAZY to adaptively guide the task planner. We show that this leads to significant speed-ups in the search for a feasible solution evaluated over unseen test environments of varying numbers of objects, goals, and initial conditions. We evaluate our TAMP approach by comparing to existing solvers for PDDLStream problems on a range of simulated 7DoF rearrangement/manipulation problems.