Search-Based Task Planning with Learned Skill Effect Models for Lifelong Robotic Manipulation

Liang, Jacky; Sharma, Mohit; Alex, LaGrassa,; Vats, Shivam; Saxena, Saumya; Kroemer, Oliver

doi:10.48550/arxiv.2109.08771

Cited by 5 publications

(11 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, in a lot of planning domains involving planning with controllers (Butzke et al 2014), sampling of states is typically not possible. One such class of planning domains where state sampling is not possible is simulator-in-the-loop planning, which uses an expensive physics simulator to generate successors (Liang et al 2021). We, therefore, focus on the more general technique of search-based planning which does not rely on state sampling.…”

Section: Parallel Sampling-based Algorithmsmentioning

confidence: 99%

ePASE: Edge-based Parallel A for Slow Evaluations

Mukherjee¹,

Aine²,

Likhachev³

2022

Preprint

View full text Add to dashboard Cite

Parallel search algorithms harness the multithreading capability of modern processors to achieve faster planning. One such algorithm is PA*SE (Parallel A* for Slow Expansions), which parallelizes state expansions to achieve faster planning in domains where state expansions are slow. In this work, we propose ePA*SE (Edge-based Parallel A* for Slow Evaluations) that improves on PA*SE by parallelizing edge evaluations instead of state expansions. This makes ePA*SE more efficient in domains where edge evaluations are expensive and need varying amounts of computational effort, which is often the case in robotics. On the theoretical front, we show that ePA*SE provides rigorous optimality guarantees. In addition, ePA*SE can be trivially extended to handle an inflation weight on the heuristic resulting in a bounded suboptimal algorithm w-ePA*SE (Weighted ePA*SE) that trades off optimality for faster planning. On the experimental front, we validate the proposed algorithm in two different planning domains: 1) motion planning for 3D humanoid navigation and 2) task and motion planning for a dual-arm robotic assembly task. We show that ePA*SE can be significantly more efficient than PA*SE and other alternatives.

show abstract

Section: Parallel Sampling-based Algorithmsmentioning

confidence: 99%

ePASE: Edge-based Parallel A for Slow Evaluations

Mukherjee¹,

Aine²,

Likhachev³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…More recent works sought to replace these components with learned modules. Notable examples include learning to predict plan feasibilities [8,15] and learning skill effect models [18,34]. While these works have relaxed many assumptions of TAMP methods, the learned modules are often built on top of handdefined representation spaces.…”

Section: A Task and Motion Planningmentioning

confidence: 99%

“…While these works have relaxed many assumptions of TAMP methods, the learned modules are often built on top of handdefined representation spaces. For example, Wang et al [33] explicitly models weight changes of a container as a result of executing a pouring skill, and Liang et al [18] constructs skill effect model based on relative object poses. As a result, these learned modules are often limited to specific tasks or domains.…”

Section: A Task and Motion Planningmentioning

confidence: 99%

“…Parameterized skills: Following prior works [1,18,36], we define a parameterized skill as a policy π ∈ Π. Each skill π is modulated by a parameter θ that specifies skillspecific execution details such as the 3D grasping poses for the grasping skill.…”

Section: A Problem Setupmentioning

confidence: 99%

“…In our task-planning formulation, there are two additional elements for a given parameterized skill: a pre-condition identifier that specifies whether the skill is executable with the given parameters θ, and a state transition model s t+1 = f T (s t , π, θ). Different from prior task planning works which either assume the pre-conditions of the skills are given [11,18] or the transition model is pre-defined [5,11], our method treats both as unknown and should be learned from experience data (Sec. III-C).…”

Section: A Problem Setupmentioning

confidence: 99%

See 2 more Smart Citations

Generalizable Task Planning through Representation Pretraining

Wang¹,

Xu²,

Li³

2022

Preprint

View full text Add to dashboard Cite

The ability to plan for multi-step manipulation tasks in unseen situations is crucial for future home robots. But collecting sufficient experience data for end-to-end learning is often infeasible in the real world, as deploying robots in many environments can be prohibitively expensive. On the other hand, large-scale scene understanding datasets contain diverse and rich semantic and geometric information. But how to leverage such information for manipulation remains an open problem. In this paper, we propose a learning-to-plan method that can generalize to new object instances by leveraging object-level representations extracted from a synthetic scene understanding dataset. We evaluate our method with a suite of challenging multi-step manipulation tasks inspired by household activities [28] and show that our model achieves measurably better success rate than state-of-the-art end-to-end approaches. Additional information can be found at: https: //sites.google.com/view/gentp

show abstract

Interleaving Graph Search and Trajectory Optimization for Aggressive Quadrotor Flight

Natarajan

Choset

Likhachev

2021

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Quadrotors can achieve aggressive flight by tracking complex manuevers and rapidly changing directions. Planning for aggressive flight with trajectory optimization could be incredibly fast, even in higher dimensions, and can account for dynamics of the quadrotor, however, only provides a locally optimal solution. On the other hand, planning with discrete graph search can handle non-convex spaces to guarantee optimality but suffers from exponential complexity with the dimension of search. We introduce a framework for aggressive quadrotor trajectory generation with global reasoning capabilities that combines the best of trajectory optimization and discrete graph search. Specifically, we develop a novel algorithmic framework that interleaves these two methods to complement each other and generate trajectories with provable guarantees on completeness up to discretization. We demonstrate and quantitatively analyze the performance of our algorithm in challenging simulation environments with narrow gaps that create severe attitude constraints and push the dynamic capabilities of the quadrotor. Experiments show the benefits of the proposed algorithmic framework over standalone trajectory optimization and graph search-based planning techniques for aggressive quadrotor flight.

show abstract

Search-Based Task Planning with Learned Skill Effect Models for Lifelong Robotic Manipulation

Cited by 5 publications

References 22 publications

ePASE: Edge-based Parallel A for Slow Evaluations

ePASE: Edge-based Parallel A for Slow Evaluations

Generalizable Task Planning through Representation Pretraining

Interleaving Graph Search and Trajectory Optimization for Aggressive Quadrotor Flight

Contact Info

Product

Resources

About

Search-Based Task Planning with Learned Skill Effect Models for Lifelong Robotic Manipulation

Cited by 5 publications

References 22 publications

ePA*SE: Edge-based Parallel A* for Slow Evaluations

ePA*SE: Edge-based Parallel A* for Slow Evaluations

Generalizable Task Planning through Representation Pretraining

Interleaving Graph Search and Trajectory Optimization for Aggressive Quadrotor Flight

Contact Info

Product

Resources

About

ePASE: Edge-based Parallel A for Slow Evaluations

ePASE: Edge-based Parallel A for Slow Evaluations