Search

Reinforcement learning-based quadruped robots excel across various terrains but still lack the ability to swim in water due to the complex underwater environment. This paper presents the development and evaluation of a data-driven hydrodynamic model for amphibious quadruped robots, aiming to enhance their adaptive capabilities in complex and dynamic underwater environments. The proposed model leverages Neural Ordinary Differential Equations (ODEs) combined with attention mechanisms to accurately process and interpret real-time sensor data. The model enables the quadruped robots to understand and predict complex environmental patterns, facilitating robust decision-making strategies. We harness real-time sensor data, capturing various environmental and internal state parameters to train and evaluate our model. A significant focus of our evaluation involves testing the quadruped robot's performance across different hydrodynamic conditions and assessing its capabilities at varying speeds and fluid dynamic conditions. The outcomes suggest that the model can effectively learn and adapt to varying conditions, enabling the prediction of force states and enhancing autonomous robotic behaviors in various practical scenarios., Comment: 8 pages, 7 figures

Large Language Models (LLMs) have been increasingly used in real-world settings, yet their strategic decision-making abilities remain largely unexplored. To fully benefit from the potential of LLMs, it's essential to understand their ability to function in complex social scenarios. Game theory, which is already used to understand real-world interactions, provides a good framework for assessing these abilities. This work investigates the performance and merits of LLMs in canonical game-theoretic two-player non-zero-sum games, Stag Hunt and Prisoner Dilemma. Our structured evaluation of GPT-3.5, GPT-4-Turbo, GPT-4o, and Llama-3-8B shows that these models, when making decisions in these games, are affected by at least one of the following systematic biases: positional bias, payoff bias, or behavioural bias. This indicates that LLMs do not fully rely on logical reasoning when making these strategic decisions. As a result, it was found that the LLMs' performance drops when the game configuration is misaligned with the affecting biases. When misaligned, GPT-3.5, GPT-4-Turbo, GPT-4o, and Llama-3-8B show an average performance drop of 32\%, 25\%, 34\%, and 29\% respectively in Stag Hunt, and 28\%, 16\%, 34\%, and 24\% respectively in Prisoner's Dilemma. Surprisingly, GPT-4o (a top-performing LLM across standard benchmarks) suffers the most substantial performance drop, suggesting that newer models are not addressing these issues. Interestingly, we found that a commonly used method of improving the reasoning capabilities of LLMs, chain-of-thought (CoT) prompting, reduces the biases in GPT-3.5, GPT-4o, and Llama-3-8B but increases the effect of the bias in GPT-4-Turbo, indicating that CoT alone cannot fully serve as a robust solution to this problem. We perform several additional experiments, which provide further insight into these observed behaviours.

Grounding natural language to the physical world is a ubiquitous topic with a wide range of applications in computer vision and robotics. Recently, 2D vision-language models such as CLIP have been widely popularized, due to their impressive capabilities for open-vocabulary grounding in 2D images. Recent works aim to elevate 2D CLIP features to 3D via feature distillation, but either learn neural fields that are scene-specific and hence lack generalization, or focus on indoor room scan data that require access to multiple camera views, which is not practical in robot manipulation scenarios. Additionally, related methods typically fuse features at pixel-level and assume that all camera views are equally informative. In this work, we show that this approach leads to sub-optimal 3D features, both in terms of grounding accuracy, as well as segmentation crispness. To alleviate this, we propose a multi-view feature fusion strategy that employs object-centric priors to eliminate uninformative views based on semantic information, and fuse features at object-level via instance segmentation masks. To distill our object-centric 3D features, we generate a large-scale synthetic multi-view dataset of cluttered tabletop scenes, spawning 15k scenes from over 3300 unique object instances, which we make publicly available. We show that our method reconstructs 3D CLIP features with improved grounding capacity and spatial consistency, while doing so from single-view RGB-D, thus departing from the assumption of multiple camera views at test time. Finally, we show that our approach can generalize to novel tabletop domains and be re-purposed for 3D instance segmentation without fine-tuning, and demonstrate its utility for language-guided robotic grasping in clutter.

Effective execution of long-horizon tasks with dexterous robotic hands remains a significant challenge in real-world problems. While learning from human demonstrations have shown encouraging results, they require extensive data collection for training. Hence, decomposing long-horizon tasks into reusable primitive skills is a more efficient approach. To achieve so, we developed DexSkills, a novel supervised learning framework that addresses long-horizon dexterous manipulation tasks using primitive skills. DexSkills is trained to recognize and replicate a select set of skills using human demonstration data, which can then segment a demonstrated long-horizon dexterous manipulation task into a sequence of primitive skills to achieve one-shot execution by the robot directly. Significantly, DexSkills operates solely on proprioceptive and tactile data, i.e., haptic data. Our real-world robotic experiments show that DexSkills can accurately segment skills, thereby enabling autonomous robot execution of a diverse range of tasks.

Scaling deep learning to huge internet-scraped datasets has yielded remarkably general capabilities in natural language processing and visual understanding and generation. In contrast, data is scarce and expensive to collect in robotics. This has seen robot learning struggle to match the generality of capabilities observed in other domains. Learning from Videos (LfV) methods seek to address this data bottleneck by augmenting traditional robot data with large internet-scraped video datasets. Such video data may provide the model with foundational information regarding physical behaviours and the physics of the world. This holds great promise for improving the generality of our robots. In this survey, we present an overview of the emerging field of LfV. We outline fundamental concepts, including the benefits and challenges of LfV. We provide a comprehensive review of current methods for: extracting knowledge from large-scale internet video; tackling key LfV challenges; and boosting downstream reinforcement and robot learning via the use of video data. LfV datasets and benchmarks are also reviewed. The survey closes with a critical discussion of challenges and opportunities. Here, we advocate for scalable foundation model approaches that can leverage the full range of available internet video to aid the learning of robot policies and dynamics models. We hope this survey can inform and catalyse further LfV research, facilitating progress towards the development of general-purpose robots., Comment: Refactored paper structure, significantly reduced paper length, rewritten abstract and introduction. Other minor improvements

To quantitatively provide reliable predictions for the hot and dense QCD matter, a holographic model should be adjusted to describe first-principles lattice results available at vanishing baryon chemical potential. The equation of state from two well-known lattice groups, the HotQCD collaboration and the Wuppertal-Budapest (WB) collaboration, shows visible differences at high temperatures. We revisit the Einstein-Maxwell-dilaton (EMD) holographic model for hot QCD with 2+1 flavors and physical quark masses by fitting lattice QCD data from the WB collaboration. Using the parameterization for the scalar potential and gauge coupling proposed in our work [Phys.Rev.D 106 (2022) 12, L121902], the equation of state, the higher order baryon number susceptibilities, and the chiral condensates are in quantitative agreement with state-of-the-art lattice results. We find that the critical endpoint (CEP) obtained from fitting the WB collaboration data is nearly identical to the one from the HotQCD collaboration, suggesting the robustness of the location of the CEP. Moreover, our holographic prediction for the CEP location is in accord with more recent Bayesian analysis on a large number of holographic EMD models and an effective potential approach of QCD from gap equations.

Recent advancements in reinforcement learning (RL) have led to remarkable achievements in robot locomotion capabilities. However, the complexity and ``black-box'' nature of neural network-based RL policies hinder their interpretability and broader acceptance, particularly in applications demanding high levels of safety and reliability. This paper introduces a novel approach to distill neural RL policies into more interpretable forms using Gradient Boosting Machines (GBMs), Explainable Boosting Machines (EBMs) and Symbolic Regression. By leveraging the inherent interpretability of generalized additive models, decision trees, and analytical expressions, we transform opaque neural network policies into more transparent ``glass-box'' models. We train expert neural network policies using RL and subsequently distill them into (i) GBMs, (ii) EBMs, and (iii) symbolic policies. To address the inherent distribution shift challenge of behavioral cloning, we propose to use the Dataset Aggregation (DAgger) algorithm with a curriculum of episode-dependent alternation of actions between expert and distilled policies, to enable efficient distillation of feedback control policies. We evaluate our approach on various robot locomotion gaits -- walking, trotting, bounding, and pacing -- and study the importance of different observations in joint actions for distilled policies using various methods. We train neural expert policies for 205 hours of simulated experience and distill interpretable policies with only 10 minutes of simulated interaction for each gait using the proposed method.

In this study, we investigate the behavior of entanglement properties in the QCD phase diagram using holographic Einstein-Maxwell-Dilaton (EMD) models. We consider two representative holographic QCD models and examine various entanglement measures, including entanglement entropy, conditional mutual information, and entanglement of purification. We find that the entanglement entropy obtained using the direct UV-cutoff renormalization method shows significant dependence on the specific model being used. However, the outcomes of other non-divergent entanglement measures, such as cutoff-independent entanglement entropy, conditional mutual information, and entanglement of purification, exhibit congruence in our calculations. When we focus specifically on the temperature range below $300$ MeV, we observe a similar behavior among the three entanglement measures. This may suggests that within this temperature range, the entanglement of QCD matter decreases with increasing temperature. Additionally, we observe distinct phase transition behaviors of entanglement properties at the critical end point (CEP) and first-order phase transitions. This observation highlights the potential of entanglement properties as effective order parameters for QCD matter phase transitions., Comment: 29 pages, 12 figures

We present a modular framework designed to enable a robot hand-arm system to learn how to catch flying objects, a task that requires fast, reactive, and accurately-timed robot motions. Our framework consists of five core modules: (i) an object state estimator that learns object trajectory prediction, (ii) a catching pose quality network that learns to score and rank object poses for catching, (iii) a reaching control policy trained to move the robot hand to pre-catch poses, (iv) a grasping control policy trained to perform soft catching motions for safe and robust grasping, and (v) a gating network trained to synthesize the actions given by the reaching and grasping policy. The former two modules are trained via supervised learning and the latter three use deep reinforcement learning in a simulated environment. We conduct extensive evaluations of our framework in simulation for each module and the integrated system, to demonstrate high success rates of in-flight catching and robustness to perturbations and sensory noise. Whilst only simple cylindrical and spherical objects are used for training, the integrated system shows successful generalization to a variety of household objects that are not used in training., Comment: 8 pages. Accepted and presented at IEEE IROS 2023

Robots operating in human-centric environments require the integration of visual grounding and grasping capabilities to effectively manipulate objects based on user instructions. This work focuses on the task of referring grasp synthesis, which predicts a grasp pose for an object referred through natural language in cluttered scenes. Existing approaches often employ multi-stage pipelines that first segment the referred object and then propose a suitable grasp, and are evaluated in private datasets or simulators that do not capture the complexity of natural indoor scenes. To address these limitations, we develop a challenging benchmark based on cluttered indoor scenes from OCID dataset, for which we generate referring expressions and connect them with 4-DoF grasp poses. Further, we propose a novel end-to-end model (CROG) that leverages the visual grounding capabilities of CLIP to learn grasp synthesis directly from image-text pairs. Our results show that vanilla integration of CLIP with pretrained models transfers poorly in our challenging benchmark, while CROG achieves significant improvements both in terms of grounding and grasping. Extensive robot experiments in both simulation and hardware demonstrate the effectiveness of our approach in challenging interactive object grasping scenarios that include clutter., Comment: Poster CoRL 2023. Dataset and code available here: https://github.com/gtziafas/OCID-VLG

We explore the phase structure of Quantum Chromodynamics (QCD) with two dynamical quark flavors at finite temperature and baryon chemical potential, employing the non-perturbative gauge/gravity duality approach. Our gravitational model is tailored to align with state-of-the-art lattice data regarding the thermal properties of multi-flavor QCD. Following a rigorous parameter calibration to match equations of state and the QCD trace anomaly at zero chemical potential derived from cutting-edge lattice QCD simulations, we investigate thermodynamic quantities and order parameters. We predict the location of the critical endpoint (CEP) at $(\mu_{\text{CEP}}, T_{\text{CEP}})=(219,182)$ MeV at which a line of first-order phase transitions terminate. We compute critical exponents associated with the CEP and find that they almost coincide with the critical exponents of the quantum 3D Ising model., Comment: 20 pages, 12 figures

Current reinforcement learning algorithms struggle in sparse and complex environments, most notably in long-horizon manipulation tasks entailing a plethora of different sequences. In this work, we propose the Intrinsically Guided Exploration from Large Language Models (IGE-LLMs) framework. By leveraging LLMs as an assistive intrinsic reward, IGE-LLMs guides the exploratory process in reinforcement learning to address intricate long-horizon with sparse rewards robotic manipulation tasks. We evaluate our framework and related intrinsic learning methods in an environment challenged with exploration, and a complex robotic manipulation task challenged by both exploration and long-horizons. Results show IGE-LLMs (i) exhibit notably higher performance over related intrinsic methods and the direct use of LLMs in decision-making, (ii) can be combined and complement existing learning methods highlighting its modularity, (iii) are fairly insensitive to different intrinsic scaling parameters, and (iv) maintain robustness against increased levels of uncertainty and horizons., Comment: Accepted at the International Conference on Robotics and Automation (ICRA), 2024. The manuscript consists of 10 pages and 6 figures

Understanding trajectory diversity is a fundamental aspect of addressing practical traffic tasks. However, capturing the diversity of trajectories presents challenges, particularly with traditional machine learning and recurrent neural networks due to the requirement of large-scale parameters. The emerging Transformer technology, renowned for its parallel computation capabilities enabling the utilization of models with hundreds of millions of parameters, offers a promising solution. In this study, we apply the Transformer architecture to traffic tasks, aiming to learn the diversity of trajectories within vehicle populations. We analyze the Transformer's attention mechanism and its adaptability to the goals of traffic tasks, and subsequently, design specific pre-training tasks. To achieve this, we create a data structure tailored to the attention mechanism and introduce a set of noises that correspond to spatio-temporal demands, which are incorporated into the structured data during the pre-training process. The designed pre-training model demonstrates excellent performance in capturing the spatial distribution of the vehicle population, with no instances of vehicle overlap and an RMSE of 0.6059 when compared to the ground truth values. In the context of time series prediction, approximately 95% of the predicted trajectories' speeds closely align with the true speeds, within a deviation of 7.5144m/s. Furthermore, in the stability test, the model exhibits robustness by continuously predicting a time series ten times longer than the input sequence, delivering smooth trajectories and showcasing diverse driving behaviors. The pre-trained model also provides a good basis for downstream fine-tuning tasks. The number of parameters of our model is over 50 million., Comment: 16 pages, 6 figures, work in update

This paper proposes a pose-graph attentional graph neural network, called P-GAT, which compares (key)nodes between sequential and non-sequential sub-graphs for place recognition tasks as opposed to a common frame-to-frame retrieval problem formulation currently implemented in SOTA place recognition methods. P-GAT uses the maximum spatial and temporal information between neighbour cloud descriptors -- generated by an existing encoder -- utilising the concept of pose-graph SLAM. Leveraging intra- and inter-attention and graph neural network, P-GAT relates point clouds captured in nearby locations in Euclidean space and their embeddings in feature space. Experimental results on the large-scale publically available datasets demonstrate the effectiveness of our approach in scenes lacking distinct features and when training and testing environments have different distributions (domain adaptation). Further, an exhaustive comparison with the state-of-the-art shows improvements in performance gains. Code is available at https://github.com/csiro-robotics/P-GAT., Comment: 10 pages, 5 figures, 5 tables, Accepted for Publication in the IEEE Robotics and Automation Letters (RA-L)

Humans can produce complex whole-body motions when interacting with their surroundings, by planning, executing and combining individual limb movements. We investigated this fundamental aspect of motor control in the setting of autonomous robotic operations. We approach this problem by hierarchical generative modelling equipped with multi-level planning-for autonomous task completion-that mimics the deep temporal architecture of human motor control. Here, temporal depth refers to the nested time scales at which successive levels of a forward or generative model unfold, for example, delivering an object requires a global plan to contextualise the fast coordination of multiple local movements of limbs. This separation of temporal scales also motivates robotics and control. Specifically, to achieve versatile sensorimotor control, it is advantageous to hierarchically structure the planning and low-level motor control of individual limbs. We use numerical and physical simulation to conduct experiments and to establish the efficacy of this formulation. Using a hierarchical generative model, we show how a humanoid robot can autonomously complete a complex task that necessitates a holistic use of locomotion, manipulation, and grasping. Specifically, we demonstrate the ability of a humanoid robot that can retrieve and transport a box, open and walk through a door to reach the destination, approach and kick a football, while showing robust performance in presence of body damage and ground irregularities. Our findings demonstrated the effectiveness of using human-inspired motor control algorithms, and our method provides a viable hierarchical architecture for the autonomous completion of challenging goal-directed tasks.

This short paper outlines two recent works on multi-contact teleoperation and the development of the SEIKO (Sequential Equilibrium Inverse Kinematic Optimization) framework. SEIKO adapts commands from the operator in real-time and ensures that the reference configuration sent to the underlying controller is feasible. Additionally, an admittance scheme is used to implement physical interaction, which is then combined with the operator's command and retargeted. SEIKO has been applied in simulations on various robots, including humanoid and quadruped robots designed for loco-manipulation. Furthermore, SEIKO has been tested on real hardware for bimanual heavy object carrying tasks., Comment: 2nd Workshop Toward Robot Avatars, 2023 IEEE International Conference on Robotics and Automation (ICRA), Jun 2023, London, United Kingdom

Traditional approaches for learning on categorical data underexploit the dependencies between columns (\aka fields) in a dataset because they rely on the embedding of data points driven alone by the classification/regression loss. In contrast, we propose a novel method for learning on categorical data with the goal of exploiting dependencies between fields. Instead of modelling statistics of features globally (i.e., by the covariance matrix of features), we learn a global field dependency matrix that captures dependencies between fields and then we refine the global field dependency matrix at the instance-wise level with different weights (so-called local dependency modelling) w.r.t. each field to improve the modelling of the field dependencies. Our algorithm exploits the meta-learning paradigm, i.e., the dependency matrices are refined in the inner loop of the meta-learning algorithm without the use of labels, whereas the outer loop intertwines the updates of the embedding matrix (the matrix performing projection) and global dependency matrix in a supervised fashion (with the use of labels). Our method is simple yet it outperforms several state-of-the-art methods on six popular dataset benchmarks. Detailed ablation studies provide additional insights into our method., Comment: IEEE Transactions on Pattern Analysis and Machine Intelligence (submitted June 2022, accepted July 2023)

Imitation learning for robot dexterous manipulation, especially with a real robot setup, typically requires a large number of demonstrations. In this paper, we present a data-efficient learning from demonstration framework which exploits the use of rich tactile sensing data and achieves fine bimanual pinch grasping. Specifically, we employ a convolutional autoencoder network that can effectively extract and encode high-dimensional tactile information. Further, We develop a framework that achieves efficient multi-sensor fusion for imitation learning, allowing the robot to learn contact-aware sensorimotor skills from demonstrations. Our comparision study against the framework without using encoded tactile features highlighted the effectiveness of incorporating rich contact information, which enabled dexterous bimanual grasping with active contact searching. Extensive experiments demonstrated the robustness of the fine pinch grasp policy directly learned from few-shot demonstration, including grasping of the same object with different initial poses, generalizing to ten unseen new objects, robust and firm grasping against external pushes, as well as contact-aware and reactive re-grasping in case of dropping objects under very large perturbations. Furthermore, the saliency map analysis method is used to describe weight distribution across various modalities during pinch grasping, confirming the effectiveness of our framework at leveraging multimodal information.

Solving long sequential tasks poses a significant challenge in embodied artificial intelligence. Enabling a robotic system to perform diverse sequential tasks with a broad range of manipulation skills is an active area of research. In this work, we present a Hybrid Hierarchical Learning framework, the Robotic Manipulation Network (ROMAN), to address the challenge of solving multiple complex tasks over long time horizons in robotic manipulation. ROMAN achieves task versatility and robust failure recovery by integrating behavioural cloning, imitation learning, and reinforcement learning. It consists of a central manipulation network that coordinates an ensemble of various neural networks, each specialising in distinct re-combinable sub-tasks to generate their correct in-sequence actions for solving complex long-horizon manipulation tasks. Experimental results show that by orchestrating and activating these specialised manipulation experts, ROMAN generates correct sequential activations for accomplishing long sequences of sophisticated manipulation tasks and achieving adaptive behaviours beyond demonstrations, while exhibiting robustness to various sensory noises. These results demonstrate the significance and versatility of ROMAN's dynamic adaptability featuring autonomous failure recovery capabilities, and highlight its potential for various autonomous manipulation tasks that demand adaptive motor skills., Comment: To appear in Nature Machine Intelligence. Includes the main and supplementary manuscript. Total of 70 pages, with a total of 9 Figures and 17 Tables

Robot motor skills can be learned through deep reinforcement learning (DRL) by neural networks as state-action mappings. While the selection of state observations is crucial, there has been a lack of quantitative analysis to date. Here, we present a systematic saliency analysis that quantitatively evaluates the relative importance of different feedback states for motor skills learned through DRL. Our approach can identify the most essential feedback states for locomotion skills, including balance recovery, trotting, bounding, pacing and galloping. By using only key states including joint positions, gravity vector, base linear and angular velocities, we demonstrate that a simulated quadruped robot can achieve robust performance in various test scenarios across these distinct skills. The benchmarks using task performance metrics show that locomotion skills learned with key states can achieve comparable performance to those with all states, and the task performance or learning success rate will drop significantly if key states are missing. This work provides quantitative insights into the relationship between state observations and specific types of motor skills, serving as a guideline for robot motor learning. The proposed method is applicable to differentiable state-action mapping, such as neural network based control policies, enabling the learning of a wide range of motor skills with minimal sensing dependencies.

Efficient traffic signal control (TSC) has been one of the most useful ways for reducing urban road congestion. Key to the challenge of TSC includes 1) the essential of real-time signal decision, 2) the complexity in traffic dynamics, and 3) the network-level coordination. Recent efforts that applied reinforcement learning (RL) methods can query policies by mapping the traffic state to the signal decision in real-time, however, is inadequate for unexpected traffic flows. By observing real traffic information, online planning methods can compute the signal decisions in a responsive manner. We propose an explicit multiagent coordination (EMC)-based online planning methods that can satisfy adaptive, real-time and network-level TSC. By multiagent, we model each intersection as an autonomous agent, and the coordination efficiency is modeled by a cost (i.e., congestion index) function between neighbor intersections. By network-level coordination, each agent exchanges messages with respect to cost function with its neighbors in a fully decentralized manner. By real-time, the message passing procedure can interrupt at any time when the real time limit is reached and agents select the optimal signal decisions according to the current message. Moreover, we prove our EMC method can guarantee network stability by borrowing ideas from transportation domain. Finally, we test our EMC method in both synthetic and real road network datasets. Experimental results are encouraging: compared to RL and conventional transportation baselines, our EMC method performs reasonably well in terms of adapting to real-time traffic dynamics, minimizing vehicle travel time and scalability to city-scale road networks.

Guaranteeing safe behaviour of reinforcement learning (RL) policies poses significant challenges for safety-critical applications, despite RL's generality and scalability. To address this, we propose a new approach to apply verification methods from control theory to learned value functions. By analyzing task structures for safety preservation, we formalize original theorems that establish links between value functions and control barrier functions. Further, we propose novel metrics for verifying value functions in safe control tasks and practical implementation details to improve learning. Our work presents a novel method for certificate learning, which unlocks a diversity of verification techniques from control theory for RL policies, and marks a significant step towards a formal framework for the general, scalable, and verifiable design of RL-based control systems. Code and videos are available at this https url: https://rl-cbf.github.io/

The cumulants of baryon number fluctuations serve as a good probe for experimentally exploring the QCD phase diagram at finite density, giving rise to characteristic fluctuation patterns associated with a possible critical endpoint (CEP). We compute the higher-order baryon number susceptibilities at finite temperature and baryon chemical potential using a holographic QCD model to address the non-perturbative aspect of strongly coupled QCD matter. The model can accurately confront lattice QCD data on a quantitative level and the location of the CEP is found to fall within the range accessible to upcoming experimental measurements. The baryon number susceptibilities up to the twelfth order are computed, and the collision energy dependence of different ratios of these susceptibilities is examined along the chemical freeze-out line. The holographic results show quantitative agreement with experimental data and the functional renormalization group results in a large collision energy range, with all ratios exhibiting a peak structure around 5-10 GeV. The mismatching between our holographic results with experimental data for sufficiently low collision energy is possibly due to non-equilibrium effects and complex experimental environments. The future experiments with measurements in the low collision energy range $\sqrt{S_{NN}}\approx 1-10~\text{GeV}$ and reduced experimental uncertainty could reveal more non-monotonic behavior signals which can be used to locate the CEP., Comment: 15 pages, 7 figures

Biological studies reveal that neural circuits located at the spinal cord called central pattern generator (CPG) oscillates and generates rhythmic signals, which are the underlying mechanism responsible for rhythmic locomotion behaviors of animals. Inspired by CPG's capability to naturally generate rhythmic patterns, researchers have attempted to create mathematical models of CPG and utilize them for the locomotion of legged robots. In this paper, we propose a network architecture that incorporates CPGs for rhythmic pattern generation and a multi-layer perceptron (MLP) network for sensory feedback. We also proposed a method that reformulates CPGs into a fully-differentiable stateless network, allowing CPGs and MLP to be jointly trained with gradient-based learning. The results show that our proposed method learned agile and dynamic locomotion policies which are capable of blind traversal over uneven terrain and resist external pushes. Simulation results also show that the learned policies are capable of self-modulating step frequency and step length to adapt to the locomotion velocity.

Continuous in-hand manipulation is an important physical interaction skill, where tactile sensing provides indispensable contact information to enable dexterous manipulation of small objects. This work proposed a framework for end-to-end policy learning with tactile feedback and sim-to-real transfer, which achieved fine in-hand manipulation that controls the pose of a thin cylindrical object, such as a long stick, to track various continuous trajectories through multiple contacts of three fingertips of a dexterous robot hand with tactile sensor arrays. We estimated the central contact position between the stick and each fingertip from the high-dimensional tactile information and showed that the learned policies achieved effective manipulation performance with the processed tactile feedback. The policies were trained with deep reinforcement learning in simulation and successfully transferred to real-world experiments, using coordinated model calibration and domain randomization. We evaluated the effectiveness of tactile information via comparative studies and validated the sim-to-real performance through real-world experiments., Comment: 10 pages, 12 figures, submitted to Transaction on Mechatronics

Videos depict the change of complex dynamical systems over time in the form of discrete image sequences. Generating controllable videos by learning the dynamical system is an important yet underexplored topic in the computer vision community. This paper presents a novel framework, TiV-ODE, to generate highly controllable videos from a static image and a text caption. Specifically, our framework leverages the ability of Neural Ordinary Differential Equations~(Neural ODEs) to represent complex dynamical systems as a set of nonlinear ordinary differential equations. The resulting framework is capable of generating videos with both desired dynamics and content. Experiments demonstrate the ability of the proposed method in generating highly controllable and visually consistent videos, and its capability of modeling dynamical systems. Overall, this work is a significant step towards developing advanced controllable video generation models that can handle complex and dynamic scenes.

Generating high-quality instance-wise grasp configurations provides critical information of how to grasp specific objects in a multi-object environment and is of high importance for robot manipulation tasks. This work proposed a novel \textbf{S}ingle-\textbf{S}tage \textbf{G}rasp (SSG) synthesis network, which performs high-quality instance-wise grasp synthesis in a single stage: instance mask and grasp configurations are generated for each object simultaneously. Our method outperforms state-of-the-art on robotic grasp prediction based on the OCID-Grasp dataset, and performs competitively on the JACQUARD dataset. The benchmarking results showed significant improvements compared to the baseline on the accuracy of generated grasp configurations. The performance of the proposed method has been validated through both extensive simulations and real robot experiments for three tasks including single object pick-and-place, grasp synthesis in cluttered environments and table cleaning task.

This work developed a kernel-based residual learning framework for quadrupedal robotic locomotion. Initially, a kernel neural network is trained with data collected from an MPC controller. Alongside a frozen kernel network, a residual controller network is trained via reinforcement learning to acquire generalized locomotion skills and resilience against external perturbations. With this proposed framework, a robust quadrupedal locomotion controller is learned with high sample efficiency and controllability, providing omnidirectional locomotion at continuous velocities. Its versatility and robustness are validated on unseen terrains that the expert MPC controller fails to traverse. Furthermore, the learned kernel can produce a range of functional locomotion behaviors and can generalize to unseen gaits.

This work developed a learning framework for perceptive legged locomotion that combines visual feedback, proprioceptive information, and active gait regulation of foot-ground contacts. The perception requires only one forward-facing camera to obtain the heightmap, and the active regulation of gait paces and traveling velocity are realized through our formulation of CPG-based high-level imitation of foot-ground contacts. Through this framework, an end-user has the ability to command task-level inputs to control different walking speeds and gait frequencies according to the traversal of different terrains, which enables more reliable negotiation with encountered obstacles. The results demonstrated that the learned perceptive locomotion policy followed task-level control inputs with intended behaviors, and was robust in presence of unseen terrains and external force perturbations. A video demonstration can be found at https://youtu.be/OTzlWzDfAe8, and the codebase at https://github.com/jennyzzt/perceptual-locomotion.