Learning a robot motor skill from scratch is impractically slow; so much so that in practice, learning must typically be bootstrapped using human demonstration. However, relying on human demonstration necessarily degrades the autonomy of robots that must learn a wide variety of skills over their operational lifetimes. We propose using kinematic motion planning as a completely autonomous, sample efficient way to bootstrap motor skill learning for object manipulation. We demonstrate the use of motion planners to bootstrap
motor skills in two complex object manipulation scenarios with different policy representations: opening a drawer with a dynamic movement primitive representation, and closing a microwave door with a deep neural network policy. We also show how our method can bootstrap a motor skill for the challenging dynamic task of learning to hit a ball off a tee, where a kinematic plan based on treating the scene as static is insufficient to solve the task, but sufficient to bootstrap a more dynamic policy. In all three cases, our method is competitive with human-demonstrated initialization, and significantly outperforms starting with a random policy. This approach enables robots to to efficiently and autonomously learn motor policies for dynamic tasks without human demonstration.
more »
« less
Learning compositional models of robot skills for task and motion planning
The objective of this work is to augment the basic abilities of a robot by learning to use sensorimotor primitives to solve complex long-horizon manipulation problems. This requires flexible generative planning that can combine primitive abilities in novel combinations and, thus, generalize across a wide variety of problems. In order to plan with primitive actions, we must have models of the actions: under what circumstances will executing this primitive successfully achieve some particular effect in the world? We use, and develop novel improvements to, state-of-the-art methods for active learning and sampling. We use Gaussian process methods for learning the constraints on skill effectiveness from small numbers of expensive-to-collect training examples. In addition, we develop efficient adaptive sampling methods for generating a comprehensive and diverse sequence of continuous candidate control parameter values (such as pouring waypoints for a cup) during planning. These values become end-effector goals for traditional motion planners that then solve for a full robot motion that performs the skill. By using learning and planning methods in conjunction, we take advantage of the strengths of each and plan for a wide variety of complex dynamic manipulation tasks. We demonstrate our approach in an integrated system, combining traditional robotics primitives with our newly learned models using an efficient robot task and motion planner. We evaluate our approach both in simulation and in the real world through measuring the quality of the selected primitive actions. Finally, we apply our integrated system to a variety of long-horizon simulated and real-world manipulation problems.
more »
« less
- Award ID(s):
- 1723381
- NSF-PAR ID:
- 10282207
- Date Published:
- Journal Name:
- The International Journal of Robotics Research
- Volume:
- 40
- Issue:
- 6-7
- ISSN:
- 0278-3649
- Page Range / eLocation ID:
- 866 to 894
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)In order to solve complex, long-horizon tasks, intelligent robots need to carry out high-level, abstract planning and reasoning in conjunction with motion planning. However, abstract models are typically lossy and plans or policies computed using them can be unexecutable. These problems are exacerbated in stochastic situations where the robot needs to reason about, and plan for multiple contingencies. We present a new approach for integrated task and motion planning in stochastic settings. In contrast to prior work in this direction, we show that our approach can effectively compute integrated task and motion policies whose branching structures encoding agent behaviors handling multiple execution-time contingencies. We prove that our algorithm is probabilistically complete and can compute feasible solution policies in an anytime fashion so that the probability of encountering an unresolved contingency decreases over time. Empirical results on a set of challenging problems show the utility and scope of our methods.more » « less
-
We present a novel method for performing integrated task and motion planning (TMP) by adapting any off-the-shelf sampling-based motion planning algorithm to simultaneously solve for a symbolically and geometrically feasible plan using a single motion planner invocation. The core insight of our technique is an embedding of symbolic state into continuous space, coupled with a novel means of automatically deriving a function guiding a planner to regions of continuous space where symbolic actions can be executed. Our technique makes few assumptions and offers a great degree of flexibility and generality compared to state of the art planners. We describe our technique and offer a proof of probabilistic completeness along with empirical evaluation of our technique on manipulation benchmark problems.more » « less
-
null (Ed.)We present a framework for planning complex motor actions such as pouring or scooping from arbitrary start states in cluttered real-world scenes. Traditional approaches to such tasks use dynamic motion primitives (DMPs) learned from human demonstrations. We enhance a recently proposed state of- the-art DMP technique capable of obstacle avoidance by including them within a novel hybrid framework. This complements DMPs with sampling-based motion planning algorithms, using the latter to explore the scene and reach promising regions from which a DMP can successfully complete the task. Experiments indicate that even obstacle-aware DMPs suffer in task success when used in scenarios which largely differ from the trained demonstration in terms of the start, goal, and obstacles. Our hybrid approach significantly outperforms obstacle-aware DMPs by successfully completing tasks in cluttered scenes for a pouring task in simulation. We further demonstrate our method on a real robot for pouring and scooping tasks.more » « less
-
more » « less
The central theme in robotic manipulation is that of the robot interacting with the world through physical contact. We tend to describe that physical contact using specific words that capture the nature of the contact and the action, such as grasp, roll, pivot, push, pull, tilt, close, open etc. We refer to these situation-specific actions as manipulation primitives. Due to the nonlinear and nonsmooth nature of physical interaction, roboticists have devoted significant efforts towards studying individual manipulation primitives. However, studying individual primitives one by one is an inherently limited process, due engineering costs, overfitting to specific tasks, and lack of robustness to unforeseen variations. These limitations motivate the main contribution of this paper: a complete and general framework to autogenerate manipulation primitives. To do so, we develop the theory and computation of contact modes as a means to classify and enumerate manipulation primitives. The contact modes form a graph, specifically a lattice. Our algorithm to autogenerate manipulation primitives (AMP) performs graph-based optimization on the contact mode lattice and solves a linear program to generate each primitive. We designed several experiments to validate our approach. We benchmarked a wide range of contact scenarios and our pipeline’s runtime was consistently in the 10 s of milliseconds. In simulation, we planned manipulation sequences using AMP. In the real-world, we showcased the robustness of our approach to real-world modeling errors. We hope that our contributions will lead to more general and robust approaches for robotic manipulation.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1177/02783649211004615
