Xiangyun Meng

Amirreza Shaban, JoonHo Lee, Sanghun Jung et al.

We introduce LiDAR-UDA, a novel two-stage self-training-based Unsupervised Domain Adaptation (UDA) method for LiDAR segmentation. Existing self-training methods use a model trained on labeled source data to generate pseudo labels for target data and refine the predictions via fine-tuning the network on the pseudo labels. These methods suffer from domain shifts caused by different LiDAR sensor configurations in the source and target domains. We propose two techniques to reduce sensor discrepancy and improve pseudo label quality: 1) LiDAR beam subsampling, which simulates different LiDAR scanning patterns by randomly dropping beams; 2) cross-frame ensembling, which exploits temporal consistency of consecutive frames to generate more reliable pseudo labels. Our method is simple, generalizable, and does not incur any extra inference cost. We evaluate our method on several public LiDAR datasets and show that it outperforms the state-of-the-art methods by more than $3.9\%$ mIoU on average for all scenarios. Code will be available at https://github.com/JHLee0513/LiDARUDA.

Yuxiang Yang, Xiangyun Meng, Wenhao Yu et al.

Jumping is essential for legged robots to traverse through difficult terrains. In this work, we propose a hierarchical framework that combines optimal control and reinforcement learning to learn continuous jumping motions for quadrupedal robots. The core of our framework is a stance controller, which combines a manually designed acceleration controller with a learned residual policy. As the acceleration controller warm starts policy for efficient training, the trained policy overcomes the limitation of the acceleration controller and improves the jumping stability. In addition, a low-level whole-body controller converts the body pose command from the stance controller to motor commands. After training in simulation, our framework can be deployed directly to the real robot, and perform versatile, continuous jumping motions, including omni-directional jumps at up to 50cm high, 60cm forward, and jump-turning at up to 90 degrees. Please visit our website for more results: https://sites.google.com/view/learning-to-jump.

Yuxiang Yang, Xiangyun Meng, Wenhao Yu et al.

The semantics of the environment, such as the terrain type and property, reveals important information for legged robots to adjust their behaviors. In this work, we present a framework that learns semantics-aware locomotion skills from perception for quadrupedal robots, such that the robot can traverse through complex offroad terrains with appropriate speeds and gaits using perception information. Due to the lack of high-fidelity outdoor simulation, our framework needs to be trained directly in the real world, which brings unique challenges in data efficiency and safety. To ensure sample efficiency, we pre-train the perception model with an off-road driving dataset. To avoid the risks of real-world policy exploration, we leverage human demonstration to train a speed policy that selects a desired forward speed from camera image. For maximum traversability, we pair the speed policy with a gait selector, which selects a robust locomotion gait for each forward speed. Using only 40 minutes of human demonstration data, our framework learns to adjust the speed and gait of the robot based on perceived terrain semantics, and enables the robot to walk over 6km without failure at close-to-optimal speed.

XiangYun Meng, XueQin Yang, Shuo Zhang

In this paper, we present the convergence analysis of the rectangular Morley element scheme utilised on the second order problem in arbitrary dimensions. Specifically, we prove that the convergence of the scheme is of $\mathcal{O}(h)$ order in energy norm and of $\mathcal{O}(h^2)$ order in $L^2$ norm on general $d$-rectangular grids. Moreover, when the grid is uniform, the convergence rate can be of $\mathcal{O}(h^2)$ order in energy norm, and the convergence rate in $L^2$ norm is still of $\mathcal{O}(h^2)$ order, which can not be improved. Numerical examples are presented to demonstrate our theoretical results.

Arthur Zhang, Xiangyun Meng, Luca Calliari et al.

Robots must adapt to diverse human instructions and operate safely in unstructured, open-world environments. Recent Vision-Language models (VLMs) offer strong priors for grounding language and perception, but remain difficult to steer for navigation due to differences in action spaces and pretraining objectives that hamper transferability to robotics tasks. Towards addressing this, we introduce VENTURA, a vision-language navigation system that finetunes internet-pretrained image diffusion models for path planning. Instead of directly predicting low-level actions, VENTURA generates a path mask (i.e. a visual plan) in image space that captures fine-grained, context-aware navigation behaviors. A lightweight behavior-cloning policy grounds these visual plans into executable trajectories, yielding an interface that follows natural language instructions to generate diverse robot behaviors. To scale training, we supervise on path masks derived from self-supervised tracking models paired with VLM-augmented captions, avoiding manual pixel-level annotation or highly engineered data collection setups. In extensive real-world evaluations, VENTURA outperforms state-of-the-art foundation model baselines on object reaching, obstacle avoidance, and terrain preference tasks, improving success rates by 33% and reducing collisions by 54% across both seen and unseen scenarios. Notably, we find that VENTURA generalizes to unseen combinations of distinct tasks, revealing emergent compositional capabilities. Videos, code, and additional materials: https://venturapath.github.io

Muhammad Fadhil Ginting, Dong-Ki Kim, Xiangyun Meng et al.

As robots become increasingly capable of operating over extended periods -- spanning days, weeks, and even months -- they are expected to accumulate knowledge of their environments and leverage this experience to assist humans more effectively. This paper studies the problem of Long-term Active Embodied Question Answering (LA-EQA), a new task in which a robot must both recall past experiences and actively explore its environment to answer complex, temporally-grounded questions. Unlike traditional EQA settings, which typically focus either on understanding the present environment alone or on recalling a single past observation, LA-EQA challenges an agent to reason over past, present, and possible future states, deciding when to explore, when to consult its memory, and when to stop gathering observations and provide a final answer. Standard EQA approaches based on large models struggle in this setting due to limited context windows, absence of persistent memory, and an inability to combine memory recall with active exploration. To address this, we propose a structured memory system for robots, inspired by the mind palace method from cognitive science. Our method encodes episodic experiences as scene-graph-based world instances, forming a reasoning and planning algorithm that enables targeted memory retrieval and guided navigation. To balance the exploration-recall trade-off, we introduce value-of-information-based stopping criteria that determines when the agent has gathered sufficient information. We evaluate our method on real-world experiments and introduce a new benchmark that spans popular simulation environments and actual industrial sites. Our approach significantly outperforms state-of-the-art baselines, yielding substantial gains in both answer accuracy and exploration efficiency.

Lirui Wang, Xiangyun Meng, Yu Xiang et al.

6D grasping in cluttered scenes is a longstanding problem in robotic manipulation. Open-loop manipulation pipelines may fail due to inaccurate state estimation, while most end-to-end grasping methods have not yet scaled to complex scenes with obstacles. In this work, we propose a new method for end-to-end learning of 6D grasping in cluttered scenes. Our hierarchical framework learns collision-free target-driven grasping based on partial point cloud observations. We learn an embedding space to encode expert grasping plans during training and a variational autoencoder to sample diverse grasping trajectories at test time. Furthermore, we train a critic network for plan selection and an option classifier for switching to an instance grasping policy through hierarchical reinforcement learning. We evaluate our method and compare against several baselines in simulation, as well as demonstrate that our latent planning can generalize to real-world cluttered-scene grasping tasks. Our videos and code can be found at https://sites.google.com/view/latent-grasping .

Xiangyun Meng, Yu Xiang, Dieter Fox

Learning high-level navigation behaviors has important implications: it enables robots to build compact visual memory for repeating demonstrations and to build sparse topological maps for planning in novel environments. Existing approaches only learn discrete, short-horizon behaviors. These standalone behaviors usually assume a discrete action space with simple robot dynamics, thus they cannot capture the intricacy and complexity of real-world trajectories. To this end, we propose Composable Behavior Embedding (CBE), a continuous behavior representation for long-horizon visual navigation. CBE is learned in an end-to-end fashion; it effectively captures path geometry and is robust to unseen obstacles. We show that CBE can be used to performing memory-efficient path following and topological mapping, saving more than an order of magnitude of memory than behavior-less approaches.

Xiangyun Meng, Nathan Ratliff, Yu Xiang et al.

Visual topological navigation has been revitalized recently thanks to the advancement of deep learning that substantially improves robot perception. However, the scalability and reliability issue remain challenging due to the complexity and ambiguity of real world images and mechanical constraints of real robots. We present an intuitive solution to show that by accurately measuring the capability of a local controller, large-scale visual topological navigation can be achieved while being scalable and robust. Our approach achieves state-of-the-art results in trajectory following and planning in large-scale environments. It also generalizes well to real robots and new environments without retraining or finetuning.

Xiangyun Meng, Nathan Ratliff, Yu Xiang et al.

End-to-end learning for autonomous navigation has received substantial attention recently as a promising method for reducing modeling error. However, its data complexity, especially around generalization to unseen environments, is high. We introduce a novel image-based autonomous navigation technique that leverages in policy structure using the Riemannian Motion Policy (RMP) framework for deep learning of vehicular control. We design a deep neural network to predict control point RMPs of the vehicle from visual images, from which the optimal control commands can be computed analytically. We show that our network trained in the Gibson environment can be used for indoor obstacle avoidance and navigation on a real RC car, and our RMP representation generalizes better to unseen environments than predicting local geometry or predicting control commands directly.