Sim-to-real transfer is a fundamental challenge in robot reinforcement learning. Discrepancies between simulation and reality can significantly impair policy performance, especially if it receives high-dimensional inputs such as dense depth estimates from vision. We propose a novel depth transfer method based on domain adaptation to bridge the visual gap between simulated and real-world depth data. A Variational Autoencoder (VAE) is first trained to encode ground-truth depth images from simulation into a latent space, which serves as input to a reinforcement learning (RL) policy. During deployment, the encoder is refined to align stereo depth images with this latent space, enabling direct policy transfer without fine-tuning. We apply our method to the task of autonomous drone navigation through cluttered environments. Experiments in IsaacGym show that our method nearly doubles the obstacle avoidance success rate when switching from ground-truth to stereo depth input. Furthermore, we demonstrate successful transfer to the photo-realistic simulator AvoidBench using only IsaacGym-generated stereo data, achieving superior performance compared to state-of-the-art baselines. Real-world evaluations in both indoor and outdoor environments confirm the effectiveness of our approach, enabling robust and generalizable depth-based navigation across diverse domains.

The humanoid robot (Unitree G1) demonstrates diverse agile whole-body skills, showcasing the control policies' agility: (a) Cristiano Ronaldo's signature celebration involving a jump with a 180-degree mid-air rotation; (b) LeBron James's "Silencer" celebration involving single-leg balancing; and (c) Kobe Bryant's famous fadeaway jump shot involving single-leg jumping and landing; (d) 1.5m-forward jumping; (e) Leg stretching; (f) 1.3m-side jumping. Abstract -- Humanoid robots hold the potential for unparalleled versatility for performing human-like, whole-body skills. However, achieving agile and coordinated whole-body motions remains a significant challenge due to the dynamics mismatch between simulation and the real world. Existing approaches, such as system identification (SysID) and domain randomization (DR) methods, often rely on labor-intensive parameter tuning or result in overly conservative policies that sacrifice agility. In this paper, we present ASAP (Aligning Simulation and Real Physics), a two-stage framework designed to tackle the dynamics mismatch and enable agile humanoid whole-body skills. Then ASAP fine-tunes pre-trained policies with the delta action model integrated into the simulator to align effectively with real-world dynamics. We evaluate ASAP across three transfer scenarios--IsaacGym to IsaacSim, IsaacGym to Genesis, and IsaacGym to the real-world Unitree G1 humanoid robot. Our approach significantly improves agility and whole-body coordination across various dynamic motions, reducing tracking error compared to SysID, DR, and delta dynamics learning baselines. ASAP enables highly agile motions that were previously difficult to achieve, demonstrating the potential of delta action learning in bridging simulation and real-world dynamics. These results suggest a promising sim-to-real direction for developing more expressive and agile humanoids. I NTRODUCTION For decades, we have envisioned humanoid robots achieving or even surpassing human-level agility. However, most prior work [46, 74, 47, 73, 107, 19, 95, 50] has primarily focused on locomotion, treating the legs as a means of mobility. Recent studies [10, 25, 24, 26, 32] have introduced whole-body expressiveness in humanoid robots, but these efforts have primarily focused on upper-body motions and have yet to achieve the agility seen in human movement.

We present a method for sampling-based model predictive control that makes use of a generic physics simulator as the dynamical model. In particular, we propose a Model Predictive Path Integral controller (MPPI), that uses the GPU-parallelizable IsaacGym simulator to compute the forward dynamics of a problem. By doing so, we eliminate the need for explicit encoding of robot dynamics and contacts with objects for MPPI. Since no explicit dynamic modeling is required, our method is easily extendable to different objects and robots and allows one to solve complex navigation and contact-rich tasks. We demonstrate the effectiveness of this method in several simulated and real-world settings, among which mobile navigation with collision avoidance, non-prehensile manipulation, and whole-body control for high-dimensional configuration spaces. This method is a powerful and accessible open-source tool to solve a large variety of contact-rich motion planning tasks.