More Like this

1. Model Error Propagation via Learned Contraction Metrics for Safe Feedback Motion Planning of Unknown Systems

   https://doi.org/10.1109/CDC45484.2021.9683354  

   Chou, Glen ; Ozay, Necmiye ; Berenson, Dmitry ( December 2021 , 2021 60th IEEE Conference on Decision and Control (CDC))

   We present a method for contraction-based feedback motion planning of locally incrementally exponentially stabilizable systems with unknown dynamics that provides probabilistic safety and reachability guarantees. Given a dynamics dataset, our method learns a deep control-affine approximation of the dynamics. To find a trusted domain where this model can be used for planning, we obtain an estimate of the Lipschitz constant of the model error, which is valid with a given probability, in a region around the training data, providing a local, spatially-varying model error bound. We derive a trajectory tracking error bound for a contraction based controller that is subjected to this model error, and then learn a controller that optimizes this tracking bound. With a given probability, we verify the correctness of the controller and tracking error bound in the trusted domain. We then use the trajectory error bound together with the trusted domain to guide a sampling-based planner to return trajectories that can be robustly tracked in execution. We show results on a 4D car, a 6D quadrotor, and a 22D deformable object manipulation task, showing our method plans safely with learned models of highdimensional underactuated systems, while baselines that plan without considering the tracking error bound ormore »the trusted domain can fail to stabilize the system and become unsafe.« less
2. Path Planning for Within-Hand Manipulation over Learned Representations of Safe States

   Calli, Berk ; Kimmel, Andrew ; Hang, Kaiyu ; Bekris, Kostas ; Dollar, Aaron ( October 2018 , International Symposium on Experimental Robotics (ISER))

   This work proposes a framework for tracking a desired path of an object held by an adaptive hand via within-hand manipulation. Such underactuated hands are able to passively achieve stable contacts with objects. Combined with vision-based control and data-driven state estimation process, they can solve tasks without accurate hand-object models or multi-modal sensory feedback. In particular, a data-driven regression process is used here to estimate the probability of dropping the object for given manipulation states. Then, an optimization-based planner aims to track the desired path while avoiding states that are above a threshold probability of dropping the object. The optimized cost function, based on the principle of Dynamic-Time Warping (DTW), seeks to minimize the area between the desired and the followed path. By adapting the threshold for the probability of dropping the object, the framework can handle objects of different weights without retraining. Experiments involving writing letters with a marker, as well as tracing randomized paths, were conducted on the Yale Model T-42 hand. Results indicate that the framework successfully avoids undesirable states, while minimizing the proposed cost function, thereby producing object paths for within-hand manipulation that closely match the target ones.
3. Hand–object configuration estimation using particle filters for dexterous in-hand manipulation

   https://doi.org/10.1177/0278364919883343  

   Hang, Kaiyu ; Bircher, Walter G. ; Morgan, Andrew S. ; Dollar, Aaron M. ( October 2019 , The International Journal of Robotics Research)

   We consider the problem of in-hand dexterous manipulation with a focus on unknown or uncertain hand–object parameters, such as hand configuration, object pose within hand, and contact positions. In particular, in this work we formulate a generic framework for hand–object configuration estimation using underactuated hands as an example. Owing to the passive reconfigurability and the lack of encoders in the hand’s joints, it is challenging to estimate, plan, and actively control underactuated manipulation. By modeling the grasp constraints, we present a particle filter-based framework to estimate the hand configuration. Specifically, given an arbitrary grasp, we start by sampling a set of hand configuration hypotheses and then randomly manipulate the object within the hand. While observing the object’s movements as evidence using an external camera, which is not necessarily calibrated with the hand frame, our estimator calculates the likelihood of each hypothesis to iteratively estimate the hand configuration. Once converged, the estimator is used to track the hand configuration in real time for future manipulations. Thereafter, we develop an algorithm to precisely plan and control the underactuated manipulation to move the grasped object to desired poses. In contrast to most other dexterous manipulation approaches, our framework does not require any tactilemore »sensing or joint encoders, and can directly operate on any novel objects, without requiring a model of the object a priori. We implemented our framework on both the Yale Model O hand and the Yale T42 hand. The results show that the estimation is accurate for different objects, and that the framework can be easily adapted across different underactuated hand models. In the end, we evaluated our planning and control algorithm with handwriting tasks, and demonstrated the effectiveness of the proposed framework.« less
4. Belief-Space Planning using Learned Models with Application to Underactuated Hands

   Kimmel, Andrew ; Sintov, Avishai ; Tan, Juntao ; Wen, Bowen ; Boularias, Abdeslam ; Bekris, Kostas E ( October 2019 , International Symposium on Robotics Research (ISRR))

   Acquiring a precise model is a challenging task for many important robotic tasks and systems - including in-hand manipulation using underactuated, adaptive hands. Learning stochastic, data-driven models is a promising alternative as they provide not only a way to propagate forward the system dynamics, but also express the uncertainty present in the collected data. Therefore, such models en- able planning in the space of state distributions, i.e., in the belief space. This paper proposes a planning framework that employs stochastic, learned models, which ex- press a distribution of states as a set of particles. The integration achieves anytime behavior in terms of returning paths of increasing quality under constraints for the probability of success to achieve a goal. The focus of this effort is on pushing the efficiency of the overall methodology despite the notorious computational hardness of belief-space planning. Experiments show that the proposed framework enables reaching a desired goal with higher success rate compared to alternatives in sim- ple benchmarks. This work also provides an application to the motivating domain of in-hand manipulation with underactuated, adaptive hands, both in the case of physically-simulated experiments as well as demonstrations with a real hand.
5. Learning compositional models of robot skills for task and motion planning

   https://doi.org/10.1177/02783649211004615  

   Wang, Zi ; Garrett, Caelan Reed ; Kaelbling, Leslie Pack ; Lozano-Pérez, Tomás ( June 2021 , The International Journal of Robotics Research)

   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 primitivesmore »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.« less