We introduce PACOH-RL, a novel model-based Meta-Reinforcement Learning (Meta-RL) algorithm designed to efficiently adapt control policies to changing dynamics. PACOH-RL meta-learns priors for the dynamics model, allowing swift adaptation to new dynamics with minimal interaction data. Existing Meta-RL methods require abundant meta-learning data, limiting their applicability in settings such as robotics, where data is costly to obtain. To address this, PACOH-RL incorporates regularization and epistemic uncertainty quantification in both the meta-learning and task adaptation stages. When facing new dynamics, we use these uncertainty estimates to effectively guide exploration and data collection. Overall, this enables positive transfer, even when access to data from prior tasks or dynamic settings is severely limited. Our experiment results demonstrate that PACOH-RL outperforms model-based RL and model-based Meta-RL baselines in adapting to new dynamic conditions. Finally, on a real robotic car, we showcase the potential for efficient RL policy adaptation in diverse, data-scarce conditions.
Collecting manipulation demonstrations with robotic hardware is tedious - and thus difficult to scale. Recording data on robot hardware ensures that it is in the appropriate format for Learning from Demonstrations (LfD) methods. By contrast, humans are proficient manipulators, and recording their actions would be easy to scale, but it is challenging to use that data format with LfD methods. The question we explore is whether there is a method to collect data in a format that can be used with LfD while retaining some of the attractive features of recording human manipulation. We propose equipping humans with hand-held, hand-actuated parallel grippers and a head-mounted camera to record demonstrations of manipulation tasks. Using customised and reproducible grippers, we collect an initial dataset of common manipulation tasks. We show that there are tasks that, against our initial intuition, can be performed using parallel grippers. Qualitative insights are obtained regarding the impact of the difference in morphology on LfD by comparing the strategies used to complete tasks with human hands and grippers. Our data collection method bridges the gap between robot- and human-native manipulation demonstration. By making the design of our gripper prototype available, we hope to reduce other researchers effort to collect manipulation data.
Nonlinear metamaterials with tailored mechanical properties have applications in engineering, medicine, robotics, and beyond. While modeling their macromechanical behavior is challenging in itself, finding structure parameters that lead to ideal approximation of high-level performance goals is a challenging task. In this work, we propose Neural Metamaterial Networks (NMN) -- smooth neural representations that encode the nonlinear mechanics of entire metamaterial families. Given structure parameters as input, NMN return continuously differentiable strain energy density functions, thus guaranteeing conservative forces by construction. Though trained on simulation data, NMN do not inherit the discontinuities resulting from topological changes in finite element meshes. They instead provide a smooth map from parameter to performance space that is fully differentiable and thus well-suited for gradient-based optimization. On this basis, we formulate inverse material design as a nonlinear programming problem that leverages neural networks for both objective functions and constraints. We use this approach to automatically design materials with desired strain-stress curves, prescribed directional stiffness and Poisson ratio profiles. We furthermore conduct ablation studies on network nonlinearities and show the advantages of our approach compared to native-scale optimization.
Precisely reconstructing and manipulating crumpled cloths is challenging due to the high dimensionality of the cloth model, as well as the limited observation at self-occluded regions. We leverage the recent progress in the field of single-view human body reconstruction to template-based reconstruct the crumpled cloths from their top-view depth observations only, with our proposed sim-real registration protocols. In contrast to previous implicit cloth representations, our reconstruction mesh explicitly indicates the positions and visibilities of the entire cloth mesh vertices, enabling more efficient dual-arm and single-arm target-oriented manipulations. Experiments demonstrate that our template-based reconstruction and target-oriented manipulation (TRTM) system can be applied to daily cloths with similar topologies as our template mesh, but have different shapes, sizes, patterns, and physical properties. Videos, datasets, pre-trained models, and code can be downloaded from our project website: https://wenbwa.github.io/TRTM/.
Unlike human beings that can employ the entire surface of their limbs as a means to establish contact with their environment, robots are typically programmed to interact with their environments via their end-effectors, in a collision-free fashion, to avoid damaging their environment. In a departure from such a traditional approach, this work presents a contact-aware controller for reference tracking that maintains interaction forces on the surface of the robot below a safety threshold in the presence of both rigid and soft contacts. Furthermore, we leveraged the proposed controller to extend the BiTRRT sample-based planning method to be contact-aware, using a simplified contact model. The effectiveness of our framework is demonstrated in hardware experiments using a Franka robot in a setup inspired by the Amazon stowing task. A demo video of our results can be seen here: https://youtu.be/2WeYytauhNg
Reinforcement learning algorithms commonly seek to optimize policies for solving one particular task. How should we explore an unknown dynamical system such that the estimated model allows us to solve multiple downstream tasks in a zero-shot manner? In this paper, we address this challenge, by developing an algorithm -- OPAX -- for active exploration. OPAX uses well-calibrated probabilistic models to quantify the epistemic uncertainty about the unknown dynamics. It optimistically -- w.r.t. to plausible dynamics -- maximizes the information gain between the unknown dynamics and state observations. We show how the resulting optimization problem can be reduced to an optimal control problem that can be solved at each episode using standard approaches. We analyze our algorithm for general models, and, in the case of Gaussian process dynamics, we give a sample complexity bound and show that the epistemic uncertainty converges to zero. In our experiments, we compare OPAX with other heuristic active exploration approaches on several environments. Our experiments show that OPAX is not only theoretically sound but also performs well for zero-shot planning on novel downstream tasks.
In this paper, we present a data-driven strategy to simplify the deployment of model-based controllers in legged robotic hardware platforms. Our approach leverages a model-free safe learning algorithm to automate the tuning of control gains, addressing the mismatch between the simplified model used in the control formulation and the real system. This method substantially mitigates the risk of hazardous interactions with the robot by sample-efficiently optimizing parameters within a probably safe region. Additionally, we extend the applicability of our approach to incorporate the different gait parameters as contexts, leading to a safe, sample-efficient exploration algorithm capable of tuning a motion controller for diverse gait patterns. We validate our method through simulation and hardware experiments, where we demonstrate that the algorithm obtains superior performance on tuning a model-based motion controller for multiple gaits safely.
This letter presents a versatile control method for dynamic and robust legged locomotion that integrates model-based optimal control with reinforcement learning (RL). Our approach involves training an RL policy to imitate reference motions generated on-demand through solving a finite-horizon optimal control problem. This integration enables the policy to leverage human expertise in generating motions to imitate while also allowing it to generalize to more complex scenarios that require a more complex dynamics model. Our method successfully learns control policies capable of generating diverse quadrupedal gait patterns and maintaining stability against unexpected external perturbations in both simulation and hardware experiments. Furthermore, we demonstrate the adaptability of our method to more complex locomotion tasks on uneven terrain without the need for excessive reward shaping or hyperparameter tuning.
Task and motion planning is one of the key problems in robotics today. It is often formulated as a discrete task allocation problem combined with continuous motion planning. Many existing approaches to TAMP involve explicit descriptions of task primitives that cause discrete changes in the kinematic relationship between the actor and the objects. In this work we propose an alternative approach to TAMP which does not involve explicit enumeration of task primitives. Instead, the actions are represented implicitly as part of the solution to a nonlinear optimization problem. We focus on decision making for robotic manipulators, specifically for pick and place tasks, and show several possible extensions. We explore the efficacy of the model through a number of simulated experiments involving multiple robots, objects and interactions with the environment.
Real-world robotic manipulation tasks remain an elusive challenge, since they involve both fine-grained environment interaction, as well as the ability to plan for long-horizon goals. Although deep reinforcement learning (RL) methods have shown encouraging results when planning end-to-end in high-dimensional environments, they remain fundamentally limited by poor sample efficiency due to inefficient exploration, and by the complexity of credit assignment over long horizons. In this work, we present Efficient Learning of High-Level Plans from Play (ELF-P), a framework for robotic learning that bridges motion planning and deep RL to achieve long-horizon complex manipulation tasks. We leverage task-agnostic play data to learn a discrete behavioral prior over object-centric primitives, modeling their feasibility given the current context. We then design a high-level goal-conditioned policy which (1) uses primitives as building blocks to scaffold complex long-horizon tasks and (2) leverages the behavioral prior to accelerate learning. We demonstrate that ELF-P has significantly better sample efficiency than relevant baselines over multiple realistic manipulation tasks and learns policies that can be easily transferred to physical hardware.