Search for: All records

Whole-body control for humanoids is challenging due to the high-dimensional nature of the problem, coupled with the inherent instability of a bipedal morphology. Learning from visual observations further exacerbates this difficulty. In this work, we explore highly data-driven approaches to visual whole-body humanoid control based on reinforcement learning, without any simplifying assumptions, reward design, or skill primitives. Specifically, we propose a hierarchical world model in which a high-level agent generates commands based on visual observations for a low-level agent to execute, both of which are trained with rewards. Our approach produces highly performant control policies in 8 tasks with a simulated 56-DoF humanoid, while synthesizing motions that are broadly preferred by humans. 

In parallel simulation, convergence and parallelism are often seen as inherently conflicting objectives. Improved parallelism typically entails lighter local computation and weaker coupling, which unavoidably slow the global convergence. This paper presents a novel GPU algorithm that achieves convergence rates comparable to fullspace Newton's method while maintaining good parallelizability just like the Jacobi method. Our approach is built on a key insight into the phenomenon ofovershoot.Overshoot occurs when a local solver aggressively minimizes its local energy without accounting for the global context, resulting in a local update that undermines global convergence. To address this, we derive a theoretically second-order optimal solution to mitigate overshoot. Furthermore, we adapt this solution into a pre-computable form. Leveraging Cubature sampling, our runtime cost is only marginally higher than the Jacobi method, yet our algorithm converges nearly quadratically as Newton's method. We also introduce a novel full-coordinate formulation for more efficient pre-computation. Our method integrates seamlessly with the incremental potential contact method and achieves second-order convergence for both stiff and soft materials. Experimental results demonstrate that our approach delivers high-quality simulations and outperforms state-of-the-art GPU methods with 50× to 100× better convergence. 

Fluid‐driven artificial muscles exhibit a behavior similar to biological muscles which makes them attractive as soft actuators for wearable assistive robots. However, state‐of‐the‐art fluidic systems typically face challenges to meet the multifaceted needs of soft wearable robots. First, soft robots are usually constrained to tethered pressure sources or bulky configurations based on flow control valves for delivery and control of high assistive forces. Second, although some soft robots exhibit untethered operation, they are significantly limited to low force capabilities. Herein, an electrohydraulic actuation system that enables both untethered and high‐force soft wearable robots is presented. This solution is achieved through a twofold design approach. First, a simplified direct‐drive actuation paradigm composed of motor, gear‐pump, and hydraulic artificial muscle (HAM) is proposed, which allows for a compact and lightweight (1.6 kg) valveless design. Second, a fluidic engine composed of a high‐torque motor with a custom‐designed gear pump is created, which is capable of generating high pressure (up to 0.75 MPa) to drive the HAM in delivering high forces (580 N). Experimental results show that the developed fluidic engine significantly outperforms state‐of‐the‐art systems in mechanical efficiency and suggest opportunities for effective deployment in soft wearable robots for human assistance. 

Abstract Shape-morphing capabilities are crucial for enabling multifunctionality in both biological and artificial systems. Various strategies for shape morphing have been proposed for applications in metamaterials and robotics. However, few of these approaches have achieved the ability to seamlessly transform into a multitude of volumetric shapes post-fabrication using a relatively simple actuation and control mechanism. Taking inspiration from thick origami and hierarchies in nature, we present a hierarchical construction method based on polyhedrons to create an extensive library of compact origami metastructures. We show that a single hierarchical origami structure can autonomously adapt to over 10³versatile architectural configurations, achieved with the utilization of fewer than 3 actuation degrees of freedom and employing simple transition kinematics. We uncover the fundamental principles governing theses shape transformation through theoretical models. Furthermore, we also demonstrate the wide-ranging potential applications of these transformable hierarchical structures. These include their uses as untethered and autonomous robotic transformers capable of various gait-shifting and multidirectional locomotion, as well as rapidly self-deployable and self-reconfigurable architecture, exemplifying its scalability up to the meter scale. Lastly, we introduce the concept of multitask reconfigurable and deployable space robots and habitats, showcasing the adaptability and versatility of these metastructures. 

Mechanical computing encodes information in deformed states of mechanical systems, such as multistable structures. However, achieving stable mechanical memory in most multistable systems remains challenging and often limited to binary information. Here, we report leveraging coupling kinematic bifurcation in rigid cube–based mechanisms with elasticity to create transformable, multistable mechanical computing metastructures with stable, high-density mechanical memory. Simply stretching the planar metastructure forms a multistable corrugated platform. It allows for independent mechanical or magnetic actuation of individual bistable element, serving as pop-up voxels for display or binary units for various tasks such as information writing, erasing, reading, encryption, and mechanologic computing. Releasing the pre-stretched strain stabilizes the prescribed information, resistant to external mechanical or magnetic perturbations, whereas re-stretching enables editable mechanical memory, akin to selective zones or disk formatting for information erasure and rewriting. Moreover, the platform can be reprogrammed and transformed into a multilayer configuration to achieve high-density memory. 

Exoskeletons have enormous potential to improve human locomotive performance. However, their development and broad dissemination are limited by the requirement for lengthy human tests and handcrafted control laws2. Here we show an experiment-free method to learn a versatile control policy in simulation. Our learning-in-simulation framework leverages dynamics-aware musculoskeletal and exoskeleton models and data-driven reinforcement learning to bridge the gap between simulation and reality without human experiments. The learned controller is deployed on a custom hip exoskeleton that automatically generates assistance across different activities with reduced metabolic rates by 24.3%, 13.1% and 15.4% for walking, running and stair climbing, respectively. Our framework may offer a generalizable and scalable strategy for the rapid development and widespread adoption of a variety of assistive robots for both able-bodied and mobility-impaired individuals.