Learning policies in simulation and transferring them to the real world has become a promising approach in dexterous manipulation. However, bridging the sim-to-real gap for each new task requires substantial human effort, such as careful reward engineering, hyperparameter tuning, and system identification. In this work, we present a system that leverages low-level skills to address these challenges for more complex tasks. Specifically, we introduce a hierarchical policy for in-hand object reorientation based on previously acquired rotation skills. This hierarchical policy learns to select which low-level skill to execute based on feedback from both the environment and the low-level skill policies themselves. Compared to learning from scratch, the hierarchical policy is more robust to out-of-distribution changes and transfers easily from simulation to real-world environments. Additionally, we propose a generalizable object pose estimator that uses proprioceptive information, low-level skill predictions, and control errors as inputs to estimate the object's pose over time. We demonstrate that our system can reorient objects, including symmetrical and textureless ones, to a desired pose.

In our daily life, we often encounter objects that are fragile and can be damaged by excessive grasping force, such as fruits. For these objects, it is paramount to grasp gently—not using the maximum amount of force possible, but rather the minimum amount of force necessary. This paper proposes using visual, tactile, and auditory signals to learn to grasp and regrasp objects stably and gently. Specifically, we use audio signals as an indicator of gentleness during the grasping, and then train an end-to-end action-conditional model from raw visuo-tactile inputs that predicts both the stability and the gentleness of future grasping candidates, thus allowing the selection and execution of the most promising action. Experimental results on a multi-fingered hand over 1,500 grasping trials demonstrated that our model is useful for gentle grasping by validating the predictive performance (3.27% higher accuracy than the vision-only variant) and providing interpretations of their behavior. Finally, real-world experiments confirmed that the grasping performance with the trained multi-modal model outperformed other baselines (17% higher rate for stable and gentle grasps than vision-only). Our approach requires neither tactile sensor calibration nor analytical force modeling, drastically reducing the engineering effort to grasp fragile objects. Dataset and videos are available at https://lasr.org/research/gentle-grasping.

Object handover is an important skill that we use daily when interacting with other humans. To deploy robots in collaborative setting, like houses, being able to receive and handing over objects safely and efficiently becomes a crucial skill. In this work, we demonstrate the use of Reinforcement Learning (RL) for dexterous object handover between two multi-finger hands. Key to this task is the use of a novel reward function based on dual quaternions to minimize the rotation distance, which outperforms other rotation representations such as Euler and rotation matrices. The robustness of the trained policy is experimentally evaluated by testing w.r.t. objects that are not included in the training distribution, and perturbations during the handover process. The results demonstrate that the trained policy successfully perform this task, achieving a total success rate of 94% in the best-case scenario after 100 experiments, thereby showing the robustness of our policy with novel objects. In addition, the best-case performance of the trained policy decreases by only 13.8% when the other robot moves during the handover, proving that our policy is also robust to this type of perturbation, which is common in real-world object handovers. Code and videos can be found here.

When robots perform complex and context-dependent tasks in our daily lives, deviations from expectations can confuse users. Explanations of the robot’s reasoning process can help users to understand the robot intentions. However, when to provide explanations and what they contain are important to avoid user annoyance. We have investigated user preferences for explanation demand and content for a robot that helps with daily cleaning tasks in a kitchen. Our results show that users want explanations in surprising situations and prefer concise explanations that clearly state the intention behind the confusing action and the contextual factors that were relevant to this decision. Based on these findings, we propose two algorithms to identify surprising actions and to construct effective explanations for Belief-Desire-Intention (BDI) robots. Our algorithms can be easily integrated in the BDI reasoning process and pave the way for better human-robot interaction with context- and user-specific explanations.

Vision-based tactile sensors have recently become popular due to their combination of low cost, very high spatial resolution, and ease of integration using widely available miniature cameras. The associated field of view and focal length, however, are difficult to package in a human-sized finger. In this article we employ optical fiber bundles to achieve a form factor that, at 15 mm diameter, is smaller than an average human fingertip. The electronics and camera are also located remotely, further reducing package size. The sensor achieves a spatial resolution of 0.22 mm and a minimum force resolution 5 mN for normal and shear contact forces. With these attributes, the DIGIT Pinki sensor is suitable for applications such as robotic and teleoperated digital palpation. We demonstrate its utility for palpation of the prostate gland and show that it can achieve clinically relevant discrimination of prostate stiffness for phantom andex vivotissue.

Touch is an important sensing modality for humans, but it has not yet been incorporated into a multimodal generative language model. This is partially due to the difficulty of obtaining natural language labels for tactile data and the complexity of aligning tactile readings with both visual observations and language descriptions. As a step towards bridging that gap, this work introduces a new dataset of 44K in-the-wild visiontouch pairs, with English language labels annotated by humans (10%) and textual pseudo-labels from GPT-4V (90%). We use this dataset to train a vision-language-aligned tactile encoder for open-vocabulary classification and a touch-visionlanguage (TVL) model for text generation using the trained encoder. Results suggest that by incorporating touch, the TVL model improves (+29% classification accuracy) tactile-vision-language alignment over existing models trained on any pair of those modalities. Although only a small fraction of the dataset is human labeled, the TVL model demonstrates improved visual-tactile understanding over GPT-4V (+12%) and open-source vision-language models (+32%) on a new touch-vision understanding benchmark. Code, checkpoints and data are available on https: //tactile-vlm.github.io.

Optical tactile sensors have recently become popular. They provide high spatial resolution, but struggle to offer fine temporal resolutions. To overcome this shortcoming, we study the idea of replacing the RGB camera with an event-based camera and introduce a new event-based optical tactile sensor called Evetac. Along with hardware design, we develop touch processing algorithms to process its measurements online at 1000 Hz. We devise an efficient algorithm to track the elastomer's deformation through the imprinted markers despite the sensor's sparse output. Benchmarking experiments demonstrate Evetac's capabilities of sensing vibrations up to 498 Hz, reconstructing shear forces, and significantly reducing data rates compared to RGB optical tactile sensors. Moreover, Evetac's output and the marker tracking provide meaningful features for learning data-driven slip detection and prediction models. The learned models form the basis for a robust and adaptive closed-loop grasp controller capable of handling a wide range of objects. We believe that fast and efficient event-based tactile sensors like Evetac will be essential for bringing human-like manipulation capabilities to robotics.

Traffic congestion in dense urban centers presents an economical and environmental burden. In recent years, the availability of vehicle- to-anything communication allows for the transmission of detailed vehicle states to the infrastructure that can be used for intelligent traffic light control. The other way around, the infrastructure can provide vehicles with advice on driving behavior, such as appropriate velocities, which can improve the efficacy of the traffic system. Several research works applied deep reinforcement learning to either traffic light control or vehicle speed advice. In this work, we propose a first attempt to jointly learn the control of both. We show this to improve the efficacy of traffic systems. In our experiments, the joint control approach reduces average vehicle trip delays, w.r.t. controlling only traffic lights, in eight out of eleven benchmark scenarios. Analyzing the qualitative behavior of the vehicle speed advice policy, we observe that this is achieved by smoothing out the velocity profile of vehicles nearby a traffic light. Learning joint control of traffic signaling and speed advice in the real world could help to reduce congestion and mitigate the economical and environmental repercussions of today's traffic systems.

As Liszt once said “(a virtuoso) must call up scent and blossom, and breathe the breath of life”, a virtuoso plays the piano with passion, poetry, and extraordinary technical ability. Hence, piano playing, being a task that is quintessentially human, becomes a hallmark for roboticians and artificial intelligence researchers to pursue. In this paper, we advocate an end-to-end reinforcement learning (RL) paradigm to demonstrate how an agent can learn directly from machine-readable music score to play the piano with touch-augmented dexterous hands on a simulated piano. To achieve the desired tasks, we design useful touch- and audio-based reward functions and a series of tasks. Empirical results show that the RL agent can not only find the correct key position but also deal with the various rhythmic, volume, and fingering requirements. As a result, the agent demonstrates its effectiveness in playing simple pieces that have different musical requirements which show the potential of leveraging reinforcement learning approach for the piano playing tasks.

The combination of Reinforcement Learning (RL) with deep learning has led to a series of impressive feats, with many believing (deep) RL provides a path towards generally capable agents. However, the success of RL agents is often highly sensitive to design choices in the training process, which may require tedious and error-prone manual tuning. This makes it challenging to use RL for new problems and also limits its full potential. In many other areas of machine learning, AutoML has shown that it is possible to automate such design choices, and AutoML has also yielded promising initial results when applied to RL. However, Automated Reinforcement Learning (AutoRL) involves not only standard applications of AutoML but also includes additional challenges unique to RL, that naturally produce a different set of methods. As such, AutoRL has been emerging as an important area of research in RL, providing promise in a variety of applications from RNA design to playing games, such as Go. Given the diversity of methods and environments considered in RL, much of the research has been conducted in distinct subfields, ranging from meta-learning to evolution. In this survey, we seek to unify the field of AutoRL, provide a common taxonomy, discuss each area in detail and pose open problems of interest to researchers going forward.

Dynamic tasks such as table tennis are relatively easy to learn for humans, but pose significant challenges to robots. Such tasks require accurate control of fast movements and precise timing in the presence of imprecise state estimation of the flying ball and the robot. Reinforcement learning (RL) has shown promise in learning complex control tasks from data. However, applying step-based RL to dynamic tasks on real systems is safety-critical as RL requires exploring and failing safely for millions of time steps in high-speed and high-acceleration regimes. This article demonstrates that using robot arms driven by pneumatic artificial muscles (PAMs) enables safe end-to-end learning of table tennis using model-free RL. In particular, we learnfrom scratchfor thousands of trials while a stochastic policy acts on the low-level controls of the real system. The robot returns andsmashesreal balls with$5 $ms−1and$12 $ms−1on average, respectively, to a desired landing point. Additionally, we present hybrid sim and real training (HYSR), a practical procedure that avoids training with real balls by virtually replaying recorded ball trajectories and applying actions to the real robot. To the best of authors’ knowledge, this work pioneers (i) failsafe learning of a safety-critical dynamic task using anthropomorphic robot arms, (ii) learning a precision-demanding problem with a PAM-driven system that is inherently hard to control as well as (iii) train a robot to play table tennis without real balls.

Model-based Reinforcement Learning (MBRL) is a promising framework for learning control in a data-efficient manner. MBRL algorithms can be fairly complex due to the separate dynamics modeling and the subsequent planning algorithm, and as a result, they often possess tens of hyperparameters and architectural choices. For this reason, MBRL typically requires significant human expertise before it can be applied to new problems and domains. To alleviate this problem, we propose to use automatic hyperparameter optimization (HPO). We demonstrate that this problem can be tackled effectively with automated HPO, which we demonstrate to yield significantly improved performance compared to human experts. In addition, we show that tuning of several MBRL hyperparameters dynamically, i.e. during the training itself, further improves the performance compared to using static hyperparameters which are kept fix for the whole training. Finally, our experiments provide valuable insights into the effects of several hyperparameters, such as plan horizon or learning rate and their influence on the stability of training and resulting rewards.

With the increased availability of rich tactile sensors, there is an an equally proportional need for open-source and integrated software capable of efficiently and effectively processing raw touch measurements into high-level signals that can be used for control and decision-making. In this paper, we present PyTouch – the first machine learning library dedicated to the processing of touch sensing signals. PyTouch, is designed to be modular, easy-to-use and provides state-of-the-art touch processing capabilities as a service with the goal of unifying the tactile sensing community by providing a library for building scalable, proven, and performance-validated modules over which applications and research can be built upon. We evaluate PyTouch on real-world data from several tactile sensors on touch processing tasks such as touch detection, slip and object pose estimations. PyTouch is open-sourced at https://github.com/facebookresearch/pytouch.

Humans build 3D understandings of the world through active object exploration, using jointly their senses of vision and touch. However, in 3D shape reconstruction, most recent progress has relied on static datasets of limited sensory data such as RGB images, depth maps or haptic readings, leaving the active exploration of the shape largely unexplored. In active touch sensing for 3D reconstruction, the goal is to actively select the tactile readings that maximize the improvement in shape reconstruction accuracy. However, the development of deep learning-based active touch models is largely limited by the lack of frameworks for shape exploration. In this paper, we focus on this problem and introduce a system composed of: 1) a haptic simulator leveraging high spatial resolution vision-based tactile sensors for active touching of 3D objects; 2) a mesh-based 3D shape reconstruction model that relies on tactile or visuotactile signals; and 3) a set of data-driven solutions with either tactile or visuotactile priors to guide the shape exploration. Our framework enables the development of the first fully data-driven solutions to active touch on top of learned models for object understanding. Our experiments show the benefits of such solutions in the task of 3D shape understanding where our models consistently outperform natural baselines. We provide our framework as a tool to foster future research in this direction.

Learning to locomote to arbitrary goals on hardware remains a challenging problem for reinforcement learning. In this paper, we present a hierarchical framework that improves sample-efficiency and generalizability of learned locomotion skills on real-world robots. Our approach divides the problem of goal-oriented locomotion into two sub-problems: learning diverse primitives skills, and using model-based planning to sequence these skills. We parametrize our primitives as cyclic movements, improving sample-efficiency of learning from scratch on a 18 degrees of freedom robot. Then, we learn coarse dynamics models over primitive cycles and use them in a model predictive control framework. This allows us to learn to walk to arbitrary goals up to 12m away, after about two hours of training from scratch on hardware. Our results on a Daisy hexapod hardware and simulation demonstrate the efficacy of our approach at reaching distant targets, in different environments, and with sensory noise.

Incorporating touch as a sensing modality for robots can enable finer and more robust manipulation skills. Existing tactile sensors are either flat, have small sensitive fields or only provide low-resolution signals. In this paper, we introduce OmniTact, a multi-directional high-resolution tactile sensor. OmniTact is designed to be used as a fingertip for robotic manipulation with robotic hands, and uses multiple micro-cameras to detect multi-directional deformations of a gel-based skin. This provides a rich signal from which a variety of different contact state variables can be inferred using modern image processing and computer vision methods. We evaluate the capabilities of OmniTact on a challenging robotic control task that requires inserting an electrical connector into an outlet, as well as a state estimation problem that is representative of those typically encountered in dexterous robotic manipulation, where the goal is to infer the angle of contact of a curved finger pressing against an object. Both tasks are performed using only touch sensing and deep convolutional neural networks to process images from the sensor's cameras. We compare with a state-of-the-art tactile sensor that is only sensitive on one side, as well as a state-of-the-art multi-directional tactile sensor, and find that OmniTact's combination of high-resolution and multi-directional sensing is crucial for reliably inserting the electrical connector and allows for higher accuracy in the state estimation task. Videos and supplementary material can be found here4.

Bayesian optimization (BO) is a popular approach to optimize expensive-to-evaluate black-box functions. A significant challenge in BO is to scale to high-dimensional parameter spaces while retaining sample efficiency. A solution considered in existing literature is to embed the high-dimensional space in a lower-dimensional manifold, often via a random linear embedding. In this paper, we identify several crucial issues and misconceptions about the use of linear embeddings for BO. We study the properties of linear embeddings from the literature and show that some of the design choices in current approaches adversely impact their performance. We show empirically that properly addressing these issues significantly improves the efficacy of linear embeddings for BO on a range of problems, including learning a gait policy for robot locomotion.

When a toddler is presented a new toy, their instinctual behaviour is to pick it up and inspect it with their hand and eyes in tandem, clearly searching over its surface to properly understand what they are playing with. At any instance here, touch provides high fidelity localized information while vision provides complementary global context. However, in 3D shape reconstruction, the complementary fusion of visual and haptic modalities remains largely unexplored. In this paper, we study this problem and present an effective chart-based approach to multi-modal shape understanding which encourages a similar fusion vision and touch information. To do so, we introduce a dataset of simulated touch and vision signals from the interaction between a robotic hand and a large array of 3D objects. Our results show that (1) leveraging both vision and touch signals consistently improves single- modality baselines; (2) our approach outperforms alternative modality fusion methods and strongly benefits from the proposed chart-based structure; (3) the reconstruction quality increases with the number of grasps provided; and (4) the touch information not only enhances the reconstruction at the touch site but also extrapolates to its local neighborhood.

Robot design is often a slow and difficult process requiring the iterative construction and testing of prototypes, with the goal of sequentially optimizing the design. For most robots, this process is further complicated by the need, when validating the capabilities of the hardware to solve the desired task, to already have an appropriate controller, which is in turn designed and tuned for the specific hardware. In this paper, we propose a novel approach, HPC-BBO, to efficiently and automatically design hardware configurations, and evaluate them by also automatically tuning the corresponding controller. HPC-BBO is based on a hierarchical Bayesian optimization process which iteratively optimizes morphology configurations (based on the performance of the previous designs during the controller learning process) and subsequently learns the corresponding controllers (exploiting the knowledge collected from optimizing for previous morphologies). Moreover, HPC-BBO can select a “batch” of multiple morphology designs at once, thus parallelizing hardware validation and reducing the number of time-consuming production cycles. We validate HPC-BBO on the design of the morphology and controller for a simulated 6-legged microrobot. Experimental results show that HPC-BBO outperforms multiple competitive baselines, and yields a 360% reduction in production cycles over standard Bayesian optimization, thus reducing the hypothetical manufacturing time of our microrobot from 21 to 4 months.

Much of the literature on robotic perception focuses on the visual modality. Vision provides a global observation of a scene, making it broadly useful. However, in the domain of robotic manipulation, vision alone can sometimes prove inadequate: in the presence of occlusions or poor lighting, visual object identification might be difficult. The sense of touch can provide robots with an alternative mechanism for recognizing objects. In this paper, we study the problem of touch-based instance recognition. We propose a novel framing of the problem as multi-modal recognition: the goal of our system is to recognize, given a visual and tactile observation, whether or not these observations correspond to the same object. To our knowledge, our work is the first to address this type of multi-modal instance recognition problem on such a large-scale with our analysis spanning 98 different objects. We employ a robot equipped with two GelSight touch sensors, one on each finger, and a self-supervised, autonomous data collection procedure to collect a dataset of tactile observations and images. Our experimental results show that it is possible to accurately recognize object instances by touch alone, including instances of novel objects that were never seen during training. Our learned model outperforms other methods on this complex task, including that of human volunteers.

Touch sensing is widely acknowledged to be important for dexterous robotic manipulation, but exploiting tactile sensing for continuous, non-prehensile manipulation is challenging. General purpose control techniques that are able to effectively leverage tactile sensing as well as accurate physics models of contacts and forces remain largely elusive, and it is unclear how to even specify a desired behavior in terms of tactile percepts. In this paper, we take a step towards addressing these issues by combining high-resolution tactile sensing with data-driven modeling using deep neural network dynamics models. We propose deep tactile MPC, a framework for learning to perform tactile servoing from raw tactile sensor inputs, without manual supervision. We show that this method enables a robot equipped with a GelSight-style tactile sensor to manipulate a ball, analog stick, and 20-sided die, learning from unsupervised autonomous interaction and then using the learned tactile predictive model to reposition each object to user-specified configurations, indicated by a goal tactile reading. Videos, visualizations and the code are available here: https://sites.google.com/view/deeptactilempc.

Model-based reinforcement learning (RL) algorithms can attain excellent sample efficiency, but often lag behind the best model-free algorithms in terms of asymptotic performance. This is especially true with high-capacity parametric function approximators, such as deep networks. In this paper, we study how to bridge this gap, by employing uncertainty-aware dynamics models. We propose a new algorithm called probabilistic ensembles with trajectory sampling (PETS) that combines uncertainty-aware deep network dynamics models with sampling-based uncertainty propagation. Our comparison to state-of-the-art model-based and model-free deep RL algorithms shows that our approach matches the asymptotic performance of model-free algorithms on several challenging benchmark tasks, while requiring significantly fewer samples (e.g. 8 and 125 times fewer samples than Soft Actor Critic and Proximal Policy Optimization respectively on the half-cheetah task).

Off-the-shelf Gaussian Process (GP) covariance functions encode smoothness assumptions on the structure of the function to be modeled. To model complex and non-differentiable functions, these smoothness assumptions are often too restrictive. One way to alleviate this limitation is to find a different representation of the data by introducing a feature space. This feature space is often learned in an unsupervised way, which might lead to data representations that are not useful for the overall regression task. In this paper, we propose Manifold Gaussian Processes, a novel supervised method that jointly learns a transformation of the data into a feature space and a GP regression from the feature space to observed space. The Manifold GP is a full GP and allows to learn data representations, which are useful for the overall regression task. As a proof-of-concept, we evaluate our approach on complex non-smooth functions where standard GPs perform poorly, such as step functions and robotics tasks with contacts.

Accurate object shape knowledge provides important information for performing stable grasping and dexterous manipulation. When modeling an object using tactile sensors, touching the object surface at a fixed grid of points can be sample inefficient. In this paper, we present an active touch strategy to efficiently reduce the surface geometry uncertainty by leveraging a probabilistic representation of object surface. In particular, we model the object surface using a Gaussian process and use the associated uncertainty information to efficiently determine the next point to explore. We validate the resulting method for tactile object surface modeling using a real robot to reconstruct multiple, complex object surfaces.

Sensor gloves are widely adopted input devices for several kinds of human-robot interaction applications. Existing glove concepts differ in features and design, but include limitations concerning the captured finger kinematics, position/orientation sensing, wireless operation, and especially economical issues. This paper presents the DAGLOVE which addresses the mentioned limitations with a low-cost design (ca. 300 €). This new sensor glove allows separate measurements of proximal and distal finger joint motions as well as position/orientation detection with an inertial measurement unit (IMU). Those sensors and tactile feedback induced by coin vibration motors at the fingertips are integrated within a wireless, easy-to-use, and open-source system. The design and implementation of hardware and software as well as proof-of-concept experiments are presented. An experimental evaluation of the sensing capabilities shows that proximal and distal finger motions can be acquired separately and that hand position/orientation can be tracked. Further, teleoperation of the iCub humanoid robot is investigated as an exemplary application to highlight the potential of the extended low-cost glove in human-robot interaction.

In whole-body control, joint torques and external forces need to be estimated accurately. In principle, this can be done through pervasive joint-torque sensing and accurate system identification. However, these sensors are expensive and may not be integrated in all links. Moreover, the exact position of the contact must be known for a precise estimation. If contacts occur on the whole body, tactile sensors can estimate the contact location, but this requires a kinematic spatial calibration, which is prone to errors. Accumulating errors may have dramatic effects on the system identification. As an alternative to classical model-based approaches we propose a data-driven mixture-of-experts learning approach using Gaussian processes. This model predicts joint torques directly from raw data of tactile and force/torque sensors. We compare our approach to an analytic model-based approach on real world data recorded from the humanoid iCub. We show that the learned model accurately predicts the joint torques resulting from contact forces, is robust to changes in the environment and outperforms existing dynamic models that use of force/ torque sensor data.

The design of gaits and corresponding control policies for bipedal walkers is a key challenge in robot locomotion. Even when a viable controller parametrization already exists, finding near-optimal parameters can be daunting. The use of automatic gait optimization methods greatly reduces the need for human expertise and time-consuming design processes. Many different approaches to automatic gait optimization have been suggested to date. However, no extensive comparison among them has yet been performed. In this paper, we present some common methods for automatic gait optimization in bipedal locomotion, and analyze their strengths and weaknesses. We experimentally evaluated these gait optimization methods on a bipedal robot, in more than 1800 experimental evaluations. In particular, we analyzed Bayesian optimization in different configurations, including various acquisition functions.

Legged locomotion is one of the most versatile forms of mobility. However, despite the importance of legged locomotion and the large number of legged robotics studies, no biped or quadruped matches the agility and versatility of their biological counterparts to date. Approaches to designing controllers for legged locomotion systems are often based on either the assumption of perfectly known dynamics or mechanical designs that substantially reduce the dimensionality of the problem. The few existing approaches for learning controllers for legged systems either require exhaustive real-world data or they improve controllers only conservatively, leading to slow learning. We present a data-efficient approach to learning feedback controllers for legged locomotive systems, based on learned probabilistic forward models for generating walking policies. On a compass walker, we show that our approach allows for learning gait policies from very little data. Moreover, we analyze learned locomotion models of a biomechanically inspired biped. Our approach has the potential to scale to high-dimensional humanoid robots with little loss in efficiency.