Abstract:Imitation Learning offers a promising approach to learn directly from data without requiring explicit models, simulations, or detailed task definitions. During inference, actions are sampled from the learned distribution and executed on the robot. However, sampled actions may fail for various reasons, and simply repeating the sampling step until a successful action is obtained can be inefficient. In this work, we propose an enhanced sampling strategy that refines the sampling distribution to avoid previously unsuccessful actions. We demonstrate that by solely utilizing data from successful demonstrations, our method can infer recovery actions without the need for additional exploratory behavior or a high-level controller. Furthermore, we leverage the concept of diffusion model decomposition to break down the primary problem (which may require long-horizon history to manage failures) into multiple smaller, more manageable sub-problems in learning, data collection, and inference, thereby enabling the system to adapt to variable failure counts. Our approach yields a low-level controller that dynamically adjusts its sampling space to improve efficiency when prior samples fall short. We validate our method across several tasks, including door opening with unknown directions, object manipulation, and button-searching scenarios, demonstrating that our approach outperforms traditional baselines.
Abstract:This research addresses the challenge of performing search missions in dynamic environments, particularly for drifting targets whose movement is dictated by a flow field. This is accomplished through a dynamical system that integrates two partial differential equations: one governing the dynamics and uncertainty of the probability distribution, and the other regulating the potential field for ergodic multi-agent search. The target probability field evolves in response to the target dynamics imposed by the environment and accomplished sensing efforts, while being explored by multiple robot agents guided by the potential field gradient. The proposed methodology was tested on two simulated search scenarios, one of which features a synthetically generated domain and showcases better performance when compared to the baseline method with static target probability over a range of agent to flow field velocity ratios. The second search scenario represents a realistic sea search and rescue mission where the search start is delayed, the search is performed in multiple robot flight missions, and the procedure for target drift uncertainty compensation is demonstrated. Furthermore, the proposed method provides an accurate survey completion metric, based on the known detection/sensing parameters, that correlates with the actual number of targets found independently.
Abstract:Generating intelligent robot behavior in contact-rich settings is a research problem where zeroth-order methods currently prevail. A major contributor to the success of such methods is their robustness in the face of non-smooth and discontinuous optimization landscapes that are characteristic of contact interactions, yet zeroth-order methods remain computationally inefficient. It is therefore desirable to develop methods for perception, planning and control in contact-rich settings that can achieve further efficiency by making use of first and second order information (i.e., gradients and Hessians). To facilitate this, we present a joint formulation of collision detection and contact modelling which, compared to existing differentiable simulation approaches, provides the following benefits: i) it results in forward and inverse dynamics that are entirely analytical (i.e. do not require solving optimization or root-finding problems with iterative methods) and smooth (i.e. twice differentiable), ii) it supports arbitrary collision geometries without needing a convex decomposition, and iii) its runtime is independent of the number of contacts. Through simulation experiments, we demonstrate the validity of the proposed formulation as a "physics for inference" that can facilitate future development of efficient methods to generate intelligent contact-rich behavior.
Abstract:Humans are experts in collaborating with others physically by regulating compliance behaviors based on the perception of their partner states and the task requirements. Enabling robots to develop proficiency in human collaboration skills can facilitate more efficient human-robot collaboration (HRC). This paper introduces an innovative impedance regulation skill learning framework for achieving HRC in multiple physical collaborative tasks. The framework is designed to adjust the robot compliance to the human partner states while adhering to reference trajectories provided by human-human demonstrations. Specifically, electromyography (EMG) signals from human muscles are collected and analyzed to extract limb impedance, representing compliance behaviors during demonstrations. Human endpoint motions are captured and represented using a probabilistic learning method to create reference trajectories and corresponding impedance profiles. Meanwhile, an LSTMbased module is implemented to develop task-oriented impedance regulation policies by mapping the muscle synergistic contributions between two demonstrators. Finally, we propose a wholebody impedance controller for a human-like robot, coordinating joint outputs to achieve the desired impedance and reference trajectory during task execution. Experimental validation was conducted through a collaborative transportation task and two interactive Tai Chi pushing hands tasks, demonstrating superior performance from the perspective of interactive forces compared to a constant impedance control method.
Abstract:We present a novel approach to enhance the performance of sampling-based Model Predictive Control (MPC) in constrained optimization by leveraging products of experts. Our methodology divides the main problem into two components: one focused on optimality and the other on feasibility. By combining the solutions from each component, represented as distributions, we apply products of experts to implement a project-then-sample strategy. In this strategy, the optimality distribution is projected into the feasible area, allowing for more efficient sampling. This approach contrasts with the traditional sample-then-project method, leading to more diverse exploration and reducing the accumulation of samples on the boundaries. We demonstrate an effective implementation of this principle using a tensor train-based distribution model, which is characterized by its non-parametric nature, ease of combination with other distributions at the task level, and straightforward sampling technique. We adapt existing tensor train models to suit this purpose and validate the efficacy of our approach through experiments in various tasks, including obstacle avoidance, non-prehensile manipulation, and tasks involving staying on manifolds. Our experimental results demonstrate that the proposed method consistently outperforms known baselines, providing strong empirical support for its effectiveness.
Abstract:Generating collision-free motions in dynamic environments is a challenging problem for high-dimensional robotics, particularly under real-time constraints. Control Barrier Functions (CBFs), widely utilized in safety-critical control, have shown significant potential for motion generation. However, for high-dimensional robot manipulators, existing QP formulations and CBF-based methods rely on positional information, overlooking higher-order derivatives such as velocities. This limitation may lead to reduced success rates, decreased performance, and inadequate safety constraints. To address this, we construct time-varying CBFs (TVCBFs) that consider velocity conditions for obstacles. Our approach leverages recent developments on distance fields for articulated manipulators, a differentiable representation that enables the mapping of objects' position and velocity into the robot's joint space, offering a comprehensive understanding of the system's interactions. This allows the manipulator to be treated as a point-mass system thus simplifying motion generation tasks. Additionally, we introduce a time-varying control Lyapunov function (TVCLF) to enable whole-body contact motions. Our approach integrates the TVCBF, TVCLF, and manipulator physical constraints within a unified QP framework. We validate our method through simulations and comparisons with state-of-the-art approaches, demonstrating its effectiveness on a 7-axis Franka robot in real-world experiments.
Abstract:Humanoid robots are envisioned as embodied intelligent agents capable of performing a wide range of human-level loco-manipulation tasks, particularly in scenarios requiring strenuous and repetitive labor. However, learning these skills is challenging due to the high degrees of freedom of humanoid robots, and collecting sufficient training data for humanoid is a laborious process. Given the rapid introduction of new humanoid platforms, a cross-embodiment framework that allows generalizable skill transfer is becoming increasingly critical. To address this, we propose a transferable framework that reduces the data bottleneck by using a unified digital human model as a common prototype and bypassing the need for re-training on every new robot platform. The model learns behavior primitives from human demonstrations through adversarial imitation, and the complex robot structures are decomposed into functional components, each trained independently and dynamically coordinated. Task generalization is achieved through a human-object interaction graph, and skills are transferred to different robots via embodiment-specific kinematic motion retargeting and dynamic fine-tuning. Our framework is validated on five humanoid robots with diverse configurations, demonstrating stable loco-manipulation and highlighting its effectiveness in reducing data requirements and increasing the efficiency of skill transfer across platforms.
Abstract:Contact-rich manipulation plays a vital role in daily human activities, yet uncertain physical parameters pose significant challenges for both model-based and model-free planning and control. A promising approach to address this challenge is to develop policies robust to a wide range of parameters. Domain adaptation and domain randomization are commonly used to achieve such policies but often compromise generalization to new instances or perform conservatively due to neglecting instance-specific information. \textit{Explicit motor adaptation} addresses these issues by estimating system parameters online and then retrieving the parameter-conditioned policy from a parameter-augmented base policy. However, it typically relies on precise system identification or additional high-quality policy retraining, presenting substantial challenges for contact-rich tasks with diverse physical parameters. In this work, we propose \textit{implicit motor adaptation}, which leverages tensor factorization as an implicit representation of the base policy. Given a roughly estimated parameter distribution, the parameter-conditioned policy can be efficiently derived by exploiting the separable structure of tensor cores from the base policy. This framework eliminates the need for precise system estimation and policy retraining while preserving optimal behavior and strong generalization. We provide a theoretical analysis validating this method, supported by numerical evaluations on three contact-rich manipulation primitives. Both simulation and real-world experiments demonstrate its ability to generate robust policies for diverse instances.
Abstract:Distance functions are crucial in robotics for representing spatial relationships between the robot and the environment. It provides an implicit representation of continuous and differentiable shapes, which can seamlessly be combined with control, optimization, and learning techniques. While standard distance fields rely on the Euclidean metric, many robotic tasks inherently involve non-Euclidean structures. To this end, we generalize the use of Euclidean distance fields to more general metric spaces by solving a Riemannian eikonal equation, a first-order partial differential equation, whose solution defines a distance field and its associated gradient flow on the manifold, enabling the computation of geodesics and globally length-minimizing paths. We show that this \emph{geodesic distance field} can also be exploited in the robot configuration space. To realize this concept, we exploit physics-informed neural networks to solve the eikonal equation for high-dimensional spaces, which provides a flexible and scalable representation without the need for discretization. Furthermore, a variant of our neural eikonal solver is introduced, which enables the gradient flow to march across both task and configuration spaces. As an example of application, we validate the proposed approach in an energy-aware motion generation task. This is achieved by considering a manifold defined by a Riemannian metric in configuration space, effectively taking the property of the robot's dynamics into account. Our approach produces minimal-energy trajectories for a 7-axis Franka robot by iteratively tracking geodesics through gradient flow backpropagation.
Abstract:Machine learning methods have a groundbreaking impact in many application domains, but their application on real robotic platforms is still limited. Despite the many challenges associated with combining machine learning technology with robotics, robot learning remains one of the most promising directions for enhancing the capabilities of robots. When deploying learning-based approaches on real robots, extra effort is required to address the challenges posed by various real-world factors. To investigate the key factors influencing real-world deployment and to encourage original solutions from different researchers, we organized the Robot Air Hockey Challenge at the NeurIPS 2023 conference. We selected the air hockey task as a benchmark, encompassing low-level robotics problems and high-level tactics. Different from other machine learning-centric benchmarks, participants need to tackle practical challenges in robotics, such as the sim-to-real gap, low-level control issues, safety problems, real-time requirements, and the limited availability of real-world data. Furthermore, we focus on a dynamic environment, removing the typical assumption of quasi-static motions of other real-world benchmarks. The competition's results show that solutions combining learning-based approaches with prior knowledge outperform those relying solely on data when real-world deployment is challenging. Our ablation study reveals which real-world factors may be overlooked when building a learning-based solution. The successful real-world air hockey deployment of best-performing agents sets the foundation for future competitions and follow-up research directions.