Planning and Execution using Inaccurate Models with Provable Guarantees

Vemula, Anirudh; Oza, Yash; Bagnell, J. Andrew; Likhachev, Maxim

doi:10.15607/rss.2020.xvi.001

Cited by 12 publications

(15 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast, we explicitly target this problem. A related problem was also addressed in (21), which made local adjustments in response to inaccurate predictions encountered in execution. This strategy is complementary to our approach of avoiding areas of predicted model inaccuracy in planning.…”

Section: Related Workmentioning

confidence: 99%

Learning where to trust unreliable models in an unstructured world for deformable object manipulation

2021

View full text Add to dashboard Cite

The world outside our laboratories seldom conforms to the assumptions of our models. This is especially true for dynamics models used in control and motion planning for complex high–degree of freedom systems like deformable objects. We must develop better models, but we must also consider that, no matter how powerful our simulators or how big our datasets, our models will sometimes be wrong. What is more, estimating how wrong models are can be difficult, because methods that predict uncertainty distributions based on training data do not account for unseen scenarios. To deploy robots in unstructured environments, we must address two key questions: When should we trust a model and what do we do if the robot is in a state where the model is unreliable. We tackle these questions in the context of planning for manipulating rope-like objects in clutter. Here, we report an approach that learns a model in an unconstrained setting and then learns a classifier to predict where that model is valid, given a limited dataset of rope-constraint interactions. We also propose a way to recover from states where our model prediction is unreliable. Our method statistically significantly outperforms learning a dynamics function and trusting it everywhere. We further demonstrate the practicality of our method on real-world mock-ups of several domestic and automotive tasks.

show abstract

Section: Related Workmentioning

confidence: 99%

Learning where to trust unreliable models in an unstructured world for deformable object manipulation

2021

View full text Add to dashboard Cite

show abstract

“…Recent works such as CMAX (Vemula et al 2020) and (McConachie et al 2020) pursue an alternative approach which does not require updating the dynamics of the model or learning a residual component. These approaches exhibit goal-driven behavior by focusing on completing the task and not on modeling the true dynamics accurately.…”

Section: Related Workmentioning

confidence: 99%

“…Following the notation of (Vemula et al 2020), we consider the deterministic shortest path problem that can be represented using the tuple M = (S, A, G, f, c) where S is the state space, A is the action space, G ⊆ S is the non-empty set of goals, f : S × A → S is a deterministic dynamics function, and c : S × A → R + ∪ {0} is the cost function. Crucially, our approach assumes that the action space A is discrete, and any goal state g ∈ G is a cost-free termination state.…”

Section: Problem Setupmentioning

confidence: 99%

CMAX++ : Leveraging Experience in Planning and Execution using Inaccurate Models

Vemula

Bagnell²,

Likhachev

2021

AAAI

Self Cite

View full text Add to dashboard Cite

Given access to accurate dynamical models, modern planning approaches are effective in computing feasible and optimal plans for repetitive robotic tasks. However, it is difficult to model the true dynamics of the real world before execution, especially for tasks requiring interactions with objects whose parameters are unknown. A recent planning approach, CMAX, tackles this problem by adapting the planner online during execution to bias the resulting plans away from inaccurately modeled regions. CMAX, while being provably guaranteed to reach the goal, requires strong assumptions on the accuracy of the model used for planning and fails to improve the quality of the solution over repetitions of the same task. In this paper we propose CMAX++, an approach that leverages real-world experience to improve the quality of resulting plans over successive repetitions of a robotic task. CMAX++ achieves this by integrating model-free learning using acquired experience with model-based planning using the potentially inaccurate model. We provide provable guarantees on the completeness and asymptotic convergence of CMAX++ to the optimal path cost as the number of repetitions increases. CMAX++ is also shown to outperform baselines in simulated robotic tasks including 3D mobile robot navigation where the track friction is incorrectly modeled, and a 7D pick-and-place task where the mass of the object is unknown leading to discrepancy between true and modeled dynamics.

show abstract

“…We can view the abstraction repair problem defined in Section III-B as a model update/repair problem. Though this is a rich area of research [1,8,21,28,30,33,38], we learn a different kind of model than those typically considered. Vemula et al [38] learns model corrections online for planning with inaccurate models, but its model updates bias the planner away from poorly modeled states rather than improving the correspondence between the model and the modeled controller.…”

Section: B Abstraction Repairmentioning

confidence: 99%

Counterexample-Guided Repair for Symbolic-Geometric Action Abstractions

Thomason,

Kress-Gazit

2021

Preprint

View full text Add to dashboard Cite

Integrated Task and Motion Planning (TMP) provides a promising class of approaches for solving robot planning problems with intricate symbolic and geometric constraints. However, the practical usefulness of TMP planners is limited by their need for symbolic abstractions of robot actions, which are difficult to construct even for experts. We propose an approach to automatically construct and continuously improve a symbolic abstraction of a robot action via observations of the robot performing the action. This approach, called automatic abstraction repair, allows symbolic abstractions to be initially incorrect or incomplete and converge toward a correct model over time. Abstraction repair uses constrained polynomial zonotopes (CPZs), an efficient non-convex set representation, to model predicates over joint symbolic and geometric state, and performs an optimizing search over symbolic edit operations to predicate formulae to improve the correspondence of a symbolic abstraction to the behavior of a physical robot controller. In this work, we describe the aforementioned predicate model, introduce the symbolic-geometric abstraction repair problem, and present an anytime algorithm for automatic abstraction repair. We then demonstrate that abstraction repair can improve realistic action abstractions for common mobile manipulation actions from a handful of observations.

show abstract

Planning and Execution using Inaccurate Models with Provable Guarantees

Cited by 12 publications

References 25 publications

Learning where to trust unreliable models in an unstructured world for deformable object manipulation

Learning where to trust unreliable models in an unstructured world for deformable object manipulation

CMAX++ : Leveraging Experience in Planning and Execution using Inaccurate Models

Counterexample-Guided Repair for Symbolic-Geometric Action Abstractions

Contact Info

Product

Resources

About