Achieving zero-shot peg insertion, where inserting an arbitrary peg into an unseen hole without task-specific training, remains a fundamental challenge in robotics. This task demands a highly generalizable perception system capable of detecting potential holes, selecting the correct mating hole from multiple candidates, estimating its precise pose, and executing insertion despite uncertainties. While learning-based methods have been applied to peg insertion, they often fail to generalize beyond the specific peg-hole pairs encountered during training. Recent advancements in Vision-Language Models (VLMs) offer a promising alternative, leveraging large-scale datasets to enable robust generalization across diverse tasks. Inspired by their success, we introduce a novel zero-shot peg insertion framework that utilizes a VLM to identify mating holes and estimate their poses without prior knowledge of their geometry. Extensive experiments demonstrate that our method achieves 90.2% accuracy, significantly outperforming baselines in identifying the correct mating hole across a wide range of previously unseen peg-hole pairs, including 3D-printed objects, toy puzzles, and industrial connectors. Furthermore, we validate the effectiveness of our approach in a real-world connector insertion task on a backpanel of a PC, where our system successfully detects holes, identifies the correct mating hole, estimates its pose, and completes the insertion with a success rate of 88.3%. These results highlight the potential of VLM-driven zero-shot reasoning for enabling robust and generalizable robotic assembly.

Object-goal navigation is a crucial engineering task for the community of embodied navigation; it involves navigating to an instance of a specified object category within unseen environments. Although extensive investigations have been conducted on both end-to-end and modular-based, data-driven approaches, fully enabling an agent to comprehend the environment through perceptual knowledge and perform object-goal navigation as efficiently as humans remains a significant challenge. Recently, large language models have shown potential in this task, thanks to their powerful capabilities for knowledge extraction and integration. In this study, we propose a data-driven, modular-based approach, trained on a dataset that incorporates common-sense knowledge of object-to-room relationships extracted from a large language model. We utilize the multi-channel Swin-Unet architecture to conduct multi-task learning incorporating with multimodal inputs. The results in the Habitat simulator demonstrate that our framework outperforms the baseline by an average of 10.6% in the efficiency metric, Success weighted by Path Length (SPL). The real-world demonstration shows that the proposed approach can efficiently conduct this task by traversing several rooms. For more details and real-world demonstrations, please check our project webpage (https://sunleyuan.github.io/ObjectNav).

Various robots, rovers, drones, and other agents of mass-produced products are expected to encounter scenes where they intersect and collaborate in the near future. In such multi-agent systems, individual identification and communication play crucial roles. In this paper, we explore camera-based visible light communication using event cameras to tackle this problem. An event camera captures the events occurring in regions with changes in brightness and can be utilized as a receiver for visible light communication, leveraging its high temporal resolution. Generally, agents with identical appearances in mass-produced products are visually indistinguishable when using conventional CMOS cameras. Therefore, linking visual information with information acquired through conventional radio communication is challenging. We empirically demonstrate the advantages of a visible light communication system employing event cameras and LEDs for visual individual identification over conventional CMOS cameras with ArUco marker recognition. In the simulation, we also verified scenarios where our event camera-based visible light communication outperforms conventional radio communication in situations with visually indistinguishable multi-agents. Finally, our newly implemented multi-agent system verifies its functionality through physical robot experiments.

The conventional few-shot classification aims at learning a model on a large labeled base dataset and rapidly adapting to a target dataset that is from the same distribution as the base dataset. However, in practice, the base and the target datasets of few-shot classification are usually from different domains, which is the problem of cross-domain few-shot classification. We tackle this problem by making a small proportion of unlabeled images in the target domain accessible in the training stage. In this setup, even though the base data are sufficient and labeled, the large domain shift still makes transferring the knowledge from the base dataset difficult. We meticulously design a cross-level knowledge distillation method, which can strengthen the ability of the model to extract more discriminative features in the target dataset by guiding the network's shallow layers to learn higher-level information. Furthermore, in order to alleviate the overfitting in the evaluation stage, we propose a feature denoising operation which can reduce the feature redundancy and mitigate overfitting. Our approach can surpass the previous state-of-the-art method, Dynamic-Distillation, by 5.44% on 1-shot and 1.37% on 5-shot classification tasks on average in the BSCD-FSL benchmark. The implementation code will be available at https://github.com/jarucezh/cldfd.

Precise perception of contact interactions is essential for the fine-grained manipulation skills for robots. In this paper, we present the design of feedback skills for robots that must learn to stack complex-shaped objects on top of each other. To design such a system, a robot should be able to reason about the stability of placement from very gentle contact interactions. Our results demonstrate that it is possible to infer the stability of object placement based on tactile readings during contact formation between the object and its environment. In particular, we estimate the contact patch between a grasped object and its environment using force and tactile observations to estimate the stability of the object during a contact formation. The contact patch could be used to estimate the stability of the object upon the release of the grasp. The proposed method is demonstrated on various pairs of objects that are used in a very popular board game.

Inverse Reinforcement Learning (IRL) is attractive in scenarios where reward engineering can be tedious. However, prior IRL algorithms use on-policy transitions, which require intensive sampling from the current policy for stable and optimal performance. This limits IRL applications in the real world, where environment interactions can become highly expensive. To tackle this problem, we present Off-Policy Inverse Reinforcement Learning (OPIRL), which (1) adopts off-policy data distribution instead of on-policy and enables significant reduction of the number of interactions with the environment, (2) learns a stationary reward function that is transferable with high generalization capabilities on changing dynamics, and (3) leverages mode-covering behavior for faster convergence. We demonstrate that our method is considerably more sample efficient and generalizes to novel environments through the experiments. Our method achieves better or comparable results on policy performance baselines with significantly fewer interactions. Furthermore, we empirically show that the recovered reward function generalizes to different tasks where prior arts are prone to fail.

The success of deep learning in the computer vision and natural language processing communities can be attributed to training of very deep neural networks with millions or billions of parameters which can then be trained with massive amounts of data. However, similar trend has largely eluded training of deep reinforcement learning (RL) algorithms where larger networks do not lead to performance improvement. Previous work has shown that this is mostly due to instability during training of deep RL agents when using larger networks. In this paper, we make an attempt to understand and address training of larger networks for deep RL. We first show that naively increasing network capacity does not improve performance. Then, we propose a novel method that consists of 1) wider networks with DenseNet connection, 2) decoupling representation learning from training of RL, 3) a distributed training method to mitigate overfitting problems. Using this three-fold technique, we show that we can train very large networks that result in significant performance gains. We present several ablation studies to demonstrate the efficacy of the proposed method and some intuitive understanding of the reasons for performance gain. We show that our proposed method outperforms other baseline algorithms on several challenging locomotion tasks.

The main novelty of the proposed approach is that it allows a robot to learn an end-to-end policy which can adapt to changes in the environment during execution. While goal conditioning of policies has been studied in the RL literature, such approaches are not easily extended to cases where the robot's goal can change during execution. This is something that humans are naturally able to do. However, it is difficult for robots to learn such reflexes (i.e., to naturally respond to dynamic environments), especially when the goal location is not explicitly provided to the robot, and instead needs to be perceived through a vision sensor. In the current work, we present a method that can achieve such behavior by combining traditional kinematic planning, deep learning, and deep reinforcement learning in a synergistic fashion to generalize to arbitrary environments. We demonstrate the proposed approach for several reaching and pick-and-place tasks in simulation, as well as on a real system of a 6-DoF industrial manipulator. A video describing our work could be found \url{https://youtu.be/hE-Ew59GRPQ}.

We present Neural A*, a novel data-driven search algorithm for path planning problems. Although data-driven planning has received much attention in recent years, little work has focused on how search-based methods can learn from demonstrations to plan better. In this work, we reformulate a canonical A* search algorithm to be differentiable and couple it with a convolutional encoder to form an end-to-end trainable neural network planner. Neural A* solves a path planning problem by (1) encoding a visual representation of the problem to estimate a movement cost map and (2) performing the A* search on the cost map to output a solution path. By minimizing the difference between the search results and ground-truth paths in demonstrations, the encoder learns to capture a variety of visual planning cues in input images, such as shapes of dead-end obstacles, bypasses, and shortcuts, which makes estimated cost maps informative. Our extensive experiments confirmed that Neural A* (a) outperformed state-of-the-art data-driven planners in terms of the search optimality and efficiency trade-off and (b) predicted realistic pedestrian paths by directly performing a search on raw image inputs.