An official website of the United States government
Generating feasible robot motions in real-time requires achieving multiple tasks (i.e., kinematic requirements) simultaneously. These tasks can have a specific goal, a range of equally valid goals, or a range of acceptable goals with a preference toward a specific goal. To satisfy multiple and potentially competing tasks simultaneously, it is important to exploit the flexibility afforded by tasks with a range of goals. In this paper, we propose a real-time motion generation method that accommodates all three categories of tasks within a single, unified framework and leverages the flexibility of tasks with a range of goals to accommodate other tasks. Our method incorporates tasks in a weighted-sum multiple-objective optimization structure and uses barrier methods with novel loss functions to encode the valid range of a task. We demonstrate the effectiveness of our method through a simulation experiment that compares it to state-of-the-art alternative approaches, and by demonstrating it on a physical camera-in-hand robot that shows that our method enables the robot to achieve smooth and feasible camera motions.
more »
« less
Composite Motion Learning with Task Control
We present a deep learning method for composite and task-driven motion control for physically simulated characters. In contrast to existing data-driven approaches using reinforcement learning that imitate full-body motions, we learn decoupled motions for specific body parts from multiple reference motions simultaneously and directly by leveraging the use of multiple discriminators in a GAN-like setup. In this process, there is no need of any manual work to produce composite reference motions for learning. Instead, the control policy explores by itself how the composite motions can be combined automatically. We further account for multiple task-specific rewards and train a single, multi-objective control policy. To this end, we propose a novel framework for multi-objective learning that adaptively balances the learning of disparate motions from multiple sources and multiple goal-directed control objectives. In addition, as composite motions are typically augmentations of simpler behaviors, we introduce a sample-efficient method for training composite control policies in an incremental manner, where we reuse a pre-trained policy as the meta policy and train a cooperative policy that adapts the meta one for new composite tasks. We show the applicability of our approach on a variety of challenging multi-objective tasks involving both composite motion imitation and multiple goal-directed control. Code is available at https://motion-lab.github.io/CompositeMotion.
more »
« less
- PAR ID:
- 10615609
- Publisher / Repository:
- ACM
- Date Published:
- Journal Name:
- ACM Transactions on Graphics
- Volume:
- 42
- Issue:
- 4
- ISSN:
- 0730-0301
- Page Range / eLocation ID:
- 1 to 16
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Performing robust goal-directed manipulation tasks remains a crucial challenge for autonomous robots. In an ideal case, shared autonomous control of manipulators would allow human users to specify their intent as a goal state and have the robot reason over the actions and motions to achieve this goal. However, realizing this goal remains elusive due to the problem of perceiving the robot’s environment. We address and describe the problem of axiomatic scene estimation for robot manipulation in cluttered scenes which is the estimation of a tree-structured scene graph describing the configuration of objects observed from robot sensing. We propose generative approaches to scene inference (as the axiomatic particle filter, and the axiomatic scene estimation by Markov chain Monte Carlo based sampler) of the robot’s environment as a scene graph. The result from AxScEs estimation are axioms amenable to goal-directed manipulation through symbolic inference for task planning and collision-free motion planning and execution. We demonstrate the results for goal-directed manipulation of multi-object scenes by a PR2 robot.more » « less
-
This paper considers the trajectory design problem for unmanned aerial vehicles (UAVs) via meta-reinforcement learning. It is assumed that the UAV can move in different directions to explore a specific area and collect data from the ground nodes (GNs) located in the area. The goal of the UAV is to reach the destination and maximize the total data collected during the flight on the trajectory while avoiding collisions with other UAVs. In the literature on UAV trajectory designs, vanilla learning algorithms are typically used to train a task-specific model, and provide near-optimal solutions for a specific spatial distribution of the GNs. However, this approach requires retraining from scratch when the locations of the GNs vary. In this work, we propose a meta reinforcement learning framework that incorporates the method of Model-Agnostic Meta-Learning (MAML). Instead of training task-specific models, we train a common initialization for different distributions of GNs and different channel conditions. From the initialization, only a few gradient descents are required for adapting to different tasks with different GN distributions and channel conditions. Additionally, we also explore when the proposed MAML framework is preferred and can outperform the compared algorithms.more » « less
-
Complex manipulation tasks often require non-trivial and coordinated movements of different parts of a robot. In this work, we address the challenges associated with learning and reproducing the skills required to execute such complex tasks. Specifically, we decompose a task into multiple subtasks and learn to reproduce the subtasks by learning stable policies from demonstrations. By leveraging the RMPflow framework for motion generation, our approach finds a stable global policy in the configuration space that enables simultaneous execution of various learned subtasks. The resulting global policy is a weighted combination of the learned policies such that the motions are coordinated and feasible under the robot's kinematic and environmental constraints. We demonstrate the necessity and efficacy of the proposed approach in the context of multiple constrained manipulation tasks performed by a Franka Emika robot.more » « less
-
Oh, A; Naumann, T; Globerson, A; Saenko, K; Hardt, M; Levine, S (Ed.)Multi-agent reinforcement learning (MARL) has primarily focused on solving a single task in isolation, while in practice the environment is often evolving, leaving many related tasks to be solved. In this paper, we investigate the benefits of meta-learning in solving multiple MARL tasks collectively. We establish the first line of theoretical results for meta-learning in a wide range of fundamental MARL settings, including learning Nash equilibria in two-player zero-sum Markov games and Markov potential games, as well as learning coarse correlated equilibria in general-sum Markov games. Under natural notions of task similarity, we show that meta-learning achieves provable sharper convergence to various game-theoretical solution concepts than learning each task separately. As an important intermediate step, we develop multiple MARL algorithms with initialization-dependent convergence guarantees. Such algorithms integrate optimistic policy mirror descents with stage-based value updates, and their refined convergence guarantees (nearly) recover the best known results even when a good initialization is unknown. To our best knowledge, such results are also new and might be of independent interest. We further provide numerical simulations to corroborate our theoretical findings.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1145/3592447
