Generative modeling of complex behaviors from labeled datasets has been a longstanding problem in decision making. Unlike language or image generation, decision making requires modeling actions - continuous-valued vectors that are multimodal in their distribution, potentially drawn from uncurated sources, where generation errors can compound in sequential prediction. A recent class of models called Behavior Transformers (BeT) addresses this by discretizing actions using k-means clustering to capture different modes. However, k-means struggles to scale for high-dimensional action spaces or long sequences, and lacks gradient information, and thus BeT suffers in modeling long-range actions. In this work, we present Vector-Quantized Behavior Transformer (VQ-BeT), a versatile model for behavior generation that handles multimodal action prediction, conditional generation, and partial observations. VQ-BeT augments BeT by tokenizing continuous actions with a hierarchical vector quantization module. Across seven environments including simulated manipulation, autonomous driving, and robotics, VQ-BeT improves on state-of-the-art models such as BeT and Diffusion Policies. Importantly, we demonstrate VQ-BeT's improved ability to capture behavior modes while accelerating inference speed 5x over Diffusion Policies. Videos and code can be found https://sjlee.cc/vq-bet
With the increasing demand for mobile robots and autonomous vehicles, several approaches for long-term robot navigation have been proposed. Among these techniques, ground segmentation and traversability estimation play important roles in perception and path planning, respectively. Even though these two techniques appear similar, their objectives are different. Ground segmentation divides data into ground and non-ground elements; thus, it is used as a preprocessing stage to extract objects of interest by rejecting ground points. In contrast, traversability estimation identifies and comprehends areas in which robots can move safely. Nevertheless, some researchers use these terms without clear distinction, leading to misunderstanding the two concepts. Therefore, in this study, we survey related literature and clearly distinguish ground and traversable regions considering four aspects: a) maneuverability of robot platforms, b) position of a robot in the surroundings, c) subset relation of negative obstacles, and d) subset relation of deformable objects.
Reinforcement learning (RL) often faces the challenges of uninformed search problems where the agent should explore without access to the domain knowledge such as characteristics of the environment or external rewards. To tackle these challenges, this work proposes a new approach for curriculum RL called Diversify for Disagreement & Conquer (D2C). Unlike previous curriculum learning methods, D2C requires only a few examples of desired outcomes and works in any environment, regardless of its geometry or the distribution of the desired outcome examples. The proposed method performs diversification of the goal-conditional classifiers to identify similarities between visited and desired outcome states and ensures that the classifiers disagree on states from out-of-distribution, which enables quantifying the unexplored region and designing an arbitrary goal-conditioned intrinsic reward signal in a simple and intuitive way. The proposed method then employs bipartite matching to define a curriculum learning objective that produces a sequence of well-adjusted intermediate goals, which enable the agent to automatically explore and conquer the unexplored region. We present experimental results demonstrating that D2C outperforms prior curriculum RL methods in both quantitative and qualitative aspects, even with the arbitrarily distributed desired outcome examples.
Recent curriculum Reinforcement Learning (RL) has shown notable progress in solving complex tasks by proposing sequences of surrogate tasks. However, the previous approaches often face challenges when they generate curriculum goals in a high-dimensional space. Thus, they usually rely on manually specified goal spaces. To alleviate this limitation and improve the scalability of the curriculum, we propose a novel curriculum method that automatically defines the semantic goal space which contains vital information for the curriculum process, and suggests curriculum goals over it. To define the semantic goal space, our method discretizes continuous observations via vector quantized-variational autoencoders (VQ-VAE) and restores the temporal relations between the discretized observations by a graph. Concurrently, ours suggests uncertainty and temporal distance-aware curriculum goals that converges to the final goals over the automatically composed goal space. We demonstrate that the proposed method allows efficient explorations in an uninformed environment with raw goal examples only. Also, ours outperforms the state-of-the-art curriculum RL methods on data efficiency and performance, in various goal-reaching tasks even with ego-centric visual inputs.
A pedestrian navigation system (PNS) in indoor environments, where global navigation satellite system (GNSS) signal access is difficult, is necessary, particularly for search and rescue (SAR) operations in large buildings. This paper focuses on studying pedestrian walking behaviors to enhance the performance of indoor pedestrian dead reckoning (PDR) and map matching techniques. Specifically, our research aims to detect pedestrian turning motions using smartphone inertial measurement unit (IMU) information in a given PDR trajectory. To improve existing methods, including the threshold-based turn detection method, hidden Markov model (HMM)-based turn detection method, and pruned exact linear time (PELT) algorithm-based turn detection method, we propose enhanced algorithms that better detect pedestrian turning motions. During field tests, using the threshold-based method, we observed a missed detection rate of 20.35% and a false alarm rate of 7.65%. The PELT-based method achieved a significant improvement with a missed detection rate of 8.93% and a false alarm rate of 6.97%. However, the best results were obtained using the HMM-based method, which demonstrated a missed detection rate of 5.14% and a false alarm rate of 2.00%. In summary, our research contributes to the development of a more accurate and reliable pedestrian navigation system by leveraging smartphone IMU data and advanced algorithms for turn detection in indoor environments.
Although many recent works have made advancements in the image restoration (IR) field, they often suffer from an excessive number of parameters. Another issue is that most Transformer-based IR methods focus only on either local or global features, leading to limited receptive fields or deficient parameter issues. To address these problems, we propose a lightweight IR network, Reciprocal Attention Mixing Transformer (RAMiT). It employs our proposed dimensional reciprocal attention mixing Transformer (D-RAMiT) blocks, which compute bi-dimensional (spatial and channel) self-attentions in parallel with different numbers of multi-heads. The bi-dimensional attentions help each other to complement their counterpart's drawbacks and are then mixed. Additionally, we introduce a hierarchical reciprocal attention mixing (H-RAMi) layer that compensates for pixel-level information losses and utilizes semantic information while maintaining an efficient hierarchical structure. Furthermore, we revisit and modify MobileNet V1 and V2 to attach efficient convolutions to our proposed components. The experimental results demonstrate that RAMiT achieves state-of-the-art performance on multiple lightweight IR tasks, including super-resolution, color denoising, grayscale denoising, low-light enhancement, and deraining. Codes will be available soon.
Current reinforcement learning (RL) often suffers when solving a challenging exploration problem where the desired outcomes or high rewards are rarely observed. Even though curriculum RL, a framework that solves complex tasks by proposing a sequence of surrogate tasks, shows reasonable results, most of the previous works still have difficulty in proposing curriculum due to the absence of a mechanism for obtaining calibrated guidance to the desired outcome state without any prior domain knowledge. To alleviate it, we propose an uncertainty & temporal distance-aware curriculum goal generation method for the outcome-directed RL via solving a bipartite matching problem. It could not only provide precisely calibrated guidance of the curriculum to the desired outcome states but also bring much better sample efficiency and geometry-agnostic curriculum goal proposal capability compared to previous curriculum RL methods. We demonstrate that our algorithm significantly outperforms these prior methods in a variety of challenging navigation tasks and robotic manipulation tasks in a quantitative and qualitative way.
As previous representations for reinforcement learning cannot effectively incorporate a human-intuitive understanding of the 3D environment, they usually suffer from sub-optimal performances. In this paper, we present Semantic-aware Neural Radiance Fields for Reinforcement Learning (SNeRL), which jointly optimizes semantic-aware neural radiance fields (NeRF) with a convolutional encoder to learn 3D-aware neural implicit representation from multi-view images. We introduce 3D semantic and distilled feature fields in parallel to the RGB radiance fields in NeRF to learn semantic and object-centric representation for reinforcement learning. SNeRL outperforms not only previous pixel-based representations but also recent 3D-aware representations both in model-free and model-based reinforcement learning.
Hierarchical Reinforcement Learning (HRL) has made notable progress in complex control tasks by leveraging temporal abstraction. However, previous HRL algorithms often suffer from serious data inefficiency as environments get large. The extended components, $i.e.$, goal space and length of episodes, impose a burden on either one or both high-level and low-level policies since both levels share the total horizon of the episode. In this paper, we present a method of Decoupling Horizons Using a Graph in Hierarchical Reinforcement Learning (DHRL) which can alleviate this problem by decoupling the horizons of high-level and low-level policies and bridging the gap between the length of both horizons using a graph. DHRL provides a freely stretchable high-level action interval, which facilitates longer temporal abstraction and faster training in complex tasks. Our method outperforms state-of-the-art HRL algorithms in typical HRL environments. Moreover, DHRL achieves long and complex locomotion and manipulation tasks.
In the field of 3D perception using 3D LiDAR sensors, ground segmentation is an essential task for various purposes, such as traversable area detection and object recognition. Under these circumstances, several ground segmentation methods have been proposed. However, some limitations are still encountered. First, some ground segmentation methods require fine-tuning of parameters depending on the surroundings, which is excessively laborious and time-consuming. Moreover, even if the parameters are well adjusted, a partial under-segmentation problem can still emerge, which implies ground segmentation failures in some regions. Finally, ground segmentation methods typically fail to estimate an appropriate ground plane when the ground is above another structure, such as a retaining wall. To address these problems, we propose a robust ground segmentation method called Patchwork++, an extension of Patchwork. Patchwork++ exploits adaptive ground likelihood estimation (A-GLE) to calculate appropriate parameters adaptively based on the previous ground segmentation results. Moreover, temporal ground revert (TGR) alleviates a partial under-segmentation problem by using the temporary ground property. Also, region-wise vertical plane fitting (R-VPF) is introduced to segment the ground plane properly even if the ground is elevated with different layers. Finally, we present reflected noise removal (RNR) to eliminate virtual noise points efficiently based on the 3D LiDAR reflection model. We demonstrate the qualitative and quantitative evaluations using a SemanticKITTI dataset. Our code is available at https://github.com/url-kaist/patchwork-plusplus