Maria Vittoria Minniti

Jinghuan Shang, Karl Schmeckpeper, Brandon B. May et al.

Vision-based robot policy learning, which maps visual inputs to actions, necessitates a holistic understanding of diverse visual tasks beyond single-task needs like classification or segmentation. Inspired by this, we introduce Theia, a vision foundation model for robot learning that distills multiple off-the-shelf vision foundation models trained on varied vision tasks. Theia's rich visual representations encode diverse visual knowledge, enhancing downstream robot learning. Extensive experiments demonstrate that Theia outperforms its teacher models and prior robot learning models using less training data and smaller model sizes. Additionally, we quantify the quality of pre-trained visual representations and hypothesize that higher entropy in feature norm distributions leads to improved robot learning performance. Code, models, and demo are available at https://theia.theaiinstitute.com.

Ran Gong, Xiaohan Zhang, Jinghuan Shang et al.

Generalist robot learning remains constrained by data: large-scale, diverse, and high-quality interaction data are expensive to collect in the real world. While simulation has become a promising way for scaling up data collection, the related tasks, including simulation task design, task-aware scene generation, expert demonstration synthesis, and sim-to-real transfer, still demand substantial human effort. We present AnyTask, an automated framework that pairs massively parallel GPU simulation with foundation models to design diverse manipulation tasks and synthesize robot data. We introduce three AnyTask agents for generating expert demonstrations aiming to solve as many tasks as possible: 1) ViPR, a novel task and motion planning agent with VLM-in-the-loop Parallel Refinement; 2) ViPR-Eureka, a reinforcement learning agent with generated dense rewards and LLM-guided contact sampling; 3) ViPR-RL, a hybrid planning and learning approach that jointly produces high-quality demonstrations with only sparse rewards. We train behavior cloning policies on generated data, validate them in simulation, and deploy them directly on real robot hardware. The policies generalize to novel object poses, achieving 44% average success across a suite of real-world pick-and-place, drawer opening, contact-rich pushing, and long-horizon manipulation tasks. Our project website is at https://anytask.rai-inst.com .

Zifan Xu, Ran Gong, Maria Vittoria Minniti et al.

Learning generalizable and robust behavior cloning policies requires large volumes of high-quality robotics data. While human demonstrations (e.g., through teleoperation) serve as the standard source for expert behaviors, acquiring such data at scale in the real world is prohibitively expensive. This paper introduces ExpertGen, a framework that automates expert policy learning in simulation to enable scalable sim-to-real transfer. ExpertGen first initializes a behavior prior using a diffusion policy trained on imperfect demonstrations, which may be synthesized by large language models or provided by humans. Reinforcement learning is then used to steer this prior toward high task success by optimizing the diffusion model's initial noise while keep original policy frozen. By keeping the pretrained diffusion policy frozen, ExpertGen regularizes exploration to remain within safe, human-like behavior manifolds, while also enabling effective learning with only sparse rewards. Empirical evaluations on challenging manipulation benchmarks demonstrate that ExpertGen reliably produces high-quality expert policies with no reward engineering. On industrial assembly tasks, ExpertGen achieves a 90.5% overall success rate, while on long-horizon manipulation tasks it attains 85% overall success, outperforming all baseline methods. The resulting policies exhibit dexterous control and remain robust across diverse initial configurations and failure states. To validate sim-to-real transfer, the learned state-based expert policies are further distilled into visuomotor policies via DAgger and successfully deployed on real robotic hardware.

Jad Abou-Chakra, Lingfeng Sun, Krishan Rana et al.

We introduce real-is-sim, a new approach to integrating simulation into behavior cloning pipelines. In contrast to real-only methods, which lack the ability to safely test policies before deployment, and sim-to-real methods, which require complex adaptation to cross the sim-to-real gap, our framework allows policies to seamlessly switch between running on real hardware and running in parallelized virtual environments. At the center of real-is-sim is a dynamic digital twin, powered by the Embodied Gaussian simulator, that synchronizes with the real world at 60Hz. This twin acts as a mediator between the behavior cloning policy and the real robot. Policies are trained using representations derived from simulator states and always act on the simulated robot, never the real one. During deployment, the real robot simply follows the simulated robot's joint states, and the simulation is continuously corrected with real world measurements. This setup, where the simulator drives all policy execution and maintains real-time synchronization with the physical world, shifts the responsibility of crossing the sim-to-real gap to the digital twin's synchronization mechanisms, instead of the policy itself. We demonstrate real-is-sim on a long-horizon manipulation task (PushT), showing that virtual evaluations are consistent with real-world results. We further show how real-world data can be augmented with virtual rollouts and compare to policies trained on different representations derived from the simulator state including object poses and rendered images from both static and robot-mounted cameras. Our results highlight the flexibility of the real-is-sim framework across training, evaluation, and deployment stages. Videos available at https://real-is-sim.github.io.

Haotian Fu, Ran Gong, Xiaohan Zhang et al.

Generalist robot policies that can perform many tasks typically require extensive expert data or simulations for training. In this work, we propose a novel Data-Efficient multitask DAgger framework that distills a single multitask policy from multiple task-specific expert policies. Our approach significantly increases the overall task success rate by actively focusing on tasks where the multitask policy underperforms. The core of our method is a performance-aware scheduling strategy that tracks how much each task's learning process benefits from the amount of data, using a Kalman filter-based estimator to robustly decide how to allocate additional demonstrations across tasks. We validate our approach on MetaWorld, as well as a suite of diverse drawer-opening tasks in IsaacLab. The resulting policy attains high performance across all tasks while using substantially fewer expert demonstrations, and the visual policy learned with our method in simulation shows better performance than naive DAgger and Behavior Cloning when transferring zero-shot to a real robot without using real data.

Maria Vittoria Minniti, Ruben Grandia, Farbod Farshidian et al.

Modern robotic systems are endowed with superior mobility and mechanical skills that make them suited to be employed in real-world scenarios, where interactions with heavy objects and precise manipulation capabilities are required. For instance, legged robots with high payload capacity can be used in disaster scenarios to remove dangerous material or carry injured people. It is thus essential to develop planning algorithms that can enable complex robots to perform motion and manipulation tasks accurately. In addition, online adaptation mechanisms with respect to new, unknown environments are needed. In this work, we impose that the optimal state-input trajectories generated by Model Predictive Control (MPC) satisfy the Lyapunov function criterion derived in adaptive control for robotic systems. As a result, we combine the stability guarantees provided by Control Lyapunov Functions (CLFs) and the optimality offered by MPC in a unified adaptive framework, yielding an improved performance during the robot's interaction with unknown objects. We validate the proposed approach in simulation and hardware tests on a quadrupedal robot carrying un-modeled payloads and pulling heavy boxes.

Johannes Pankert, Maria Vittoria Minniti, Lorenz Wellhausen et al.

Measurement update rules for Bayes filters often contain hand-crafted heuristics to compute observation probabilities for high-dimensional sensor data, like images. In this work, we propose the novel approach Deep Measurement Update (DMU) as a general update rule for a wide range of systems. DMU has a conditional encoder-decoder neural network structure to process depth images as raw inputs. Even though the network is trained only on synthetic data, the model shows good performance at evaluation time on real-world data. With our proposed training scheme primed data training , we demonstrate how the DMU models can be trained efficiently to be sensitive to condition variables without having to rely on a stochastic information bottleneck. We validate the proposed methods in multiple scenarios of increasing complexity, beginning with the pose estimation of a single object to the joint estimation of the pose and the internal state of an articulated system. Moreover, we provide a benchmark against Articulated Signed Distance Functions(A-SDF) on the RBO dataset as a baseline comparison for articulation state estimation.

Mattia Risiglione, Jean-Pierre Sleiman, Maria Vittoria Minniti et al.

When dealing with the haptic teleoperation of multi-limbed mobile manipulators, the problem of mitigating the destabilizing effects arising from the communication link between the haptic device and the remote robot has not been properly addressed. In this work, we propose a passive control architecture to haptically teleoperate a legged mobile manipulator, while remaining stable in the presence of time delays and frequency mismatches in the master and slave controllers. At the master side, a discrete-time energy modulation of the control input is proposed. At the slave side, passivity constraints are included in an optimization-based whole-body controller to satisfy the energy limitations. A hybrid teleoperation scheme allows the human operator to remotely operate the robot's end-effector while in stance mode, and its base velocity in locomotion mode. The resulting control architecture is demonstrated on a quadrupedal robot with an artificial delay added to the network.

Maria Vittoria Minniti, Ruben Grandia, Kevin Fäh et al.

Modern, torque-controlled service robots can regulate contact forces when interacting with their environment. Model Predictive Control (MPC) is a powerful method to solve the underlying control problem, allowing to plan for whole-body motions while including different constraints imposed by the robot dynamics or its environment. However, an accurate model of the robot-environment is needed to achieve a satisfying closed-loop performance. Currently, this necessity undermines the performance and generality of MPC in manipulation tasks. In this work, we combine an MPC-based whole-body controller with two adaptive schemes, derived from online system identification and adaptive control. As a result, we enable a general mobile manipulator to interact with unknown environments, without any need for re-tuning parameters or pre-modeling the interacting objects. In combination with the MPC controller, the two adaptive approaches are validated and benchmarked with a ball-balancing manipulator in door opening and object lifting tasks.

Jean-Pierre Sleiman, Farbod Farshidian, Maria Vittoria Minniti et al.

In this paper, we propose a whole-body planning framework that unifies dynamic locomotion and manipulation tasks by formulating a single multi-contact optimal control problem. We model the hybrid nature of a generic multi-limbed mobile manipulator as a switched system, and introduce a set of constraints that can encode any pre-defined gait sequence or manipulation schedule in the formulation. Since the system is designed to actively manipulate its environment, the equations of motion are composed by augmenting the robot's centroidal dynamics with the manipulated-object dynamics. This allows us to describe any high-level task in the same cost/constraint function. The resulting planning framework could be solved on the robot's onboard computer in real-time within a model predictive control scheme. This is demonstrated in a set of real hardware experiments done in free-motion, such as base or end-effector pose tracking, and while pushing/pulling a heavy resistive door. Robustness against model mismatches and external disturbances is also verified during these test cases.

Maria Vittoria Minniti, Farbod Farshidian, Ruben Grandia et al.

Autonomous mobile manipulation offers a dual advantage of mobility provided by a mobile platform and dexterity afforded by the manipulator. In this paper, we present a whole-body optimal control framework to jointly solve the problems of manipulation, balancing and interaction as one optimization problem for an inherently unstable robot. The optimization is performed using a Model Predictive Control (MPC) approach; the optimal control problem is transcribed at the end-effector space, treating the position and orientation tasks in the MPC planner, and skillfully planning for end-effector contact forces. The proposed formulation evaluates how the control decisions aimed at end-effector tracking and environment interaction will affect the balance of the system in the future. We showcase the advantages of the proposed MPC approach on the example of a ball-balancing robot with a robotic manipulator and validate our controller in hardware experiments for tasks such as end-effector pose tracking and door opening.